High School Art & Design

Between 1892 and 1894 Claude Monet painted the west front of Rouen Cathedral more than thirty times. He rented rooms across the square, set up several canvases at once, and moved from one to the next as the light changed, then finished the group in his studio at Giverny. The Metropolitan Museum of Art holds one of them, Rouen Cathedral: The Portal (Sunlight), dated 1894. Nobody thinks Monet painted the same picture thirty times because he could not get it right. He was not painting a cathedral. He was painting what light does to stone, and one canvas cannot hold an answer to that.

That is a sustained investigation, and it is the organising idea of this whole course. You are not going to make eighteen unrelated pieces. You are going to find one question and keep pushing on it, so that at the end you have a body of work with an argument inside it.

A folder against a body of work

Most first-year portfolios look like this: a still life of fruit, a self-portrait, a copy of an album cover, an abstract thing in acrylic, a charcoal hand. Each one may be competent. Together they say nothing, because there is no thread. A viewer moving from piece to piece learns only that you can use different materials.

A body of work looks different. The pieces talk to each other. A decision made in the third one gets tested in the seventh. Something fails in the fifth and gets solved in the ninth. The viewer can see you thinking across the sequence. That visible thinking is the thing being assessed by every serious art programme and the thing that makes a portfolio worth looking at.

	A folder of pieces	A body of work
Held together by	Nothing, or a medium	A question
Order	Arbitrary; any sequence works	Developmental; later work depends on earlier work
Repetition	Looks like running out of ideas	Looks like testing a variable
Failure	Hidden or thrown away	Kept, because it explains the next decision
What a viewer learns	What you can do	What you are trying to find out

Key idea: a body of work is not a set of pieces that look alike. It is a set of pieces that are asking the same question with different variables changed.

Three artists, three variables

Look at what actually changes across a real series, because the discipline is in the variable.

1. Monet, Rouen Cathedral, 1892 to 1894. The subject is fixed to the point of stubbornness: the same facade, from almost the same window. The variable is light, and therefore colour. Because everything else is nailed down, the series can say something precise about how a grey stone wall becomes blue, pink and gold. If he had also changed the building, the series would prove nothing.
2. Hokusai, Thirty-six Views of Mount Fuji, about 1830 to 1832. The mountain is the constant. The variable is viewpoint and context: Fuji seen past a wave, behind scaffolding, through a barrel, over rooftops. The Met holds the most famous sheet in the series, Under the Wave off Kanagawa, in which the mountain is the smallest thing in the picture. Same subject, thirty-six different arguments about what a mountain is for.
3. Vija Celmins, from the 1960s onward. Celmins has spent decades on a very short list of subjects: ocean surfaces, night skies, spider webs, all rendered in dense graphite or paint, all without a horizon. Tate holds work of hers. The variable here is not light or viewpoint but attention and surface, and the series works precisely because the subjects refuse to develop.

Notice that none of these is a theme in the school sense. "The ocean" is not an investigation. "What happens to an image when there is nothing in it to give you scale" is.

Inquiry, practice, evidence

A sustained investigation has three parts, and students usually have one of them.

Part	What it is	What it looks like when it is missing
Inquiry	The question you are asking, stated in words you wrote yourself	Work that is skilful and about nothing
Practice	The making: repeated attempts, materials tested, things tried and dropped	One finished piece per idea, no attempts in between
Evidence	The record: sketchbook, photographs of stages, notes on what changed and why	A portfolio that cannot explain itself without you standing next to it

This is not an invention of this course. The College Board's AP Art and Design portfolios are built on exactly this structure: a Sustained Investigation section of fifteen images, submitted with written responses describing the inquiry and how practice, experimentation and revision moved it forward, alongside a smaller Selected Works section chosen for quality. Whether or not you ever sit that exam, the framework is a good description of what a year of serious art-making looks like.

What a question looks like

The difference between a topic and a question is that a question can be answered wrongly. Compare:

Topic (weak)	Question (workable)
My family	Can a room describe a person who is not in it?
Nature	What happens to a plant drawing when I remove every line and keep only the shadow?
Identity	How much of a face can I take away before it stops being a particular person?
The city	What does a place look like drawn only from the sounds I heard standing in it?
Anxiety	Can a mark made fast read as a different emotion from the same mark made slowly?

Each question on the right names a thing you can change and a result you can look at. That is what makes it generative: you can do it fifteen times and get fifteen different answers. The topics on the left generate one picture each and then stop.

Do not worry about whether the question is profound. Profundity is not the criterion; productivity is. A narrow, slightly odd question about shadows will carry a year of work better than a large statement about identity.

Repetition is not the same as iteration

Making the same drawing twice is repetition and teaches you almost nothing after the first time. Making it again with one thing deliberately changed is iteration, and it is how every one of the artists above worked.

Pick the variable before you start. Some that work: scale, medium, light source, viewpoint, speed of mark, colour range, cropping, surface, whether the subject is present or drawn from memory. Change one at a time. If you change three at once and the result is better, you will not know which change did it.

The point: iteration is an experiment with a control. Hold everything fixed except the thing you want to learn about.

Keep the failures

The strongest evidence in a portfolio is often the piece that did not work, sitting next to the piece that fixed it. It shows a reader the decision you made, and a decision is the only thing that distinguishes an artist from a person with good hands.

So photograph work in progress. Keep the version you abandoned. Write one sentence in your sketchbook when something goes wrong: not "this is bad" but "the background is competing with the head, so next time I will drop its value range". That sentence is worth more later than the drawing was.

How this course is organised

Modules 2 to 5 build the technical range you will need: observational drawing, value, perspective, colour, composition, mark-making, printmaking, collage, photography and digital tools. Module 6 handles the things that decide whether the work travels: using sources responsibly, documenting properly, writing about your own work, and running a critique.

Do not wait until you feel skilful to start the investigation. Choose the question in the next lesson and let it run underneath everything else. The technical lessons are more useful when you have somewhere to put them.

One honest limit: this course cannot see your work. It can tell you what to look at, what to change, and what questions to ask yourself in front of a drawing, and Lesson 18 gives you a critique protocol you can run with other people. But no page can stand behind you and say "your ellipse is tilting". Find one other person, a teacher, a class, an online group, and show them work regularly. Feedback is the part of art education that has to come from outside your own head.

Common misconceptions

• "A body of work means everything has to look the same." It means everything is driven by the same inquiry. Monet's cathedrals look startlingly different from one another; that difference is the content.
• "I should pick the question and never change it." Questions get sharper as you work. What you must not do is change the question every week, or replace it silently and pretend the earlier work was always about the new thing. Record the change and why it happened.
• "Process work is padding to fill the sketchbook." It is the evidence. A portfolio without it asks a viewer to take the finished work on trust.
• "Only original ideas count." Nobody expects a seventeen year old to invent a new question in art. They expect you to ask a real one and pursue it honestly, which is rarer than it sounds.
• "Skill comes first, then ideas." They grow together. Skill acquired with no question attached tends to plateau into competent, forgettable work.

Where this leaves us

A sustained investigation is a body of work generated by a question, developed through iteration, and documented so that the thinking is visible. Monet held the cathedral still and changed the light; Hokusai held Fuji still and changed the viewpoint; Celmins holds almost everything still and changes the quality of attention. All three demonstrate the same discipline: fix the constants, change one variable, and let the series do the arguing. The three components you owe your portfolio are an inquiry you can state, a practice with visible attempts and revisions, and evidence of both. Keep the failures, because they carry the decisions. In the next lesson you will write the question you are going to live with, and test whether it can generate fifteen different pieces.

Sources

1. Monet, C. (1894). Rouen Cathedral: The Portal (Sunlight) [Oil on canvas]. The Metropolitan Museum of Art, New York. metmuseum.org
2. Hokusai, K. (ca. 1830-32). Under the Wave off Kanagawa, from Thirty-six Views of Mount Fuji [Woodblock print]. The Metropolitan Museum of Art, New York. metmuseum.org
3. Tate. Vija Celmins: artist page and works in the collection. Tate, London. tate.org.uk
4. College Board. AP Art and Design: course and exam description, sustained investigation. AP Central. apcentral.collegeboard.org

Jacob Lawrence was twenty-three when he began The Migration Series in 1940: sixty small panels in tempera on hardboard, telling the movement of Black Americans out of the rural South. He did not paint them one at a time. He laid out all sixty panels at once and worked colour by colour across the entire set, so that every black in the series went down before any other colour did. The Museum of Modern Art in New York holds the even-numbered panels and the Phillips Collection in Washington holds the odd-numbered ones, which is why the series is almost never seen whole.

That working method only makes sense if you already know what the whole series is about before you begin. Which is the problem this lesson solves: where a question that can carry sixty panels, or fifteen, actually comes from.

Questions do not come from a list of themes

If someone hands you "identity, conflict, nature, memory" and asks you to pick, you will produce illustration: a picture about a word. Real inquiries come from somewhere much more local. Six places to look, in rough order of how often they work:

1. Something you keep looking at anyway. Not what you think is worthy. What your camera roll is actually full of.
2. A constraint you find interesting. One colour. One line weight. Only what fits in your hand. Constraints generate work faster than freedom does.
3. A material that behaves oddly. Ink on wet paper. Torn newsprint. A photocopier making a copy of a copy of a copy.
4. A place you can return to. A hundred metres of pavement. One room. One window at different hours, which is what Monet did.
5. A piece of evidence. A box of family photographs, a school report, a receipt, an object nobody else in the house cares about.
6. Something that annoys you. Irritation is a reliable engine and is much easier to sustain than admiration.

Why this matters: a question you can get to on foot will outlast a question you had to look up.

Four tests

Before you commit, run any candidate through these. A question that fails one is usually rescuable; a question that fails three is a topic.

Test	Ask yourself	Fails if
Generative	Can this produce fifteen genuinely different pieces?	You can only picture two, and the second is the first one bigger
Specific	Could an attempt at this be wrong?	Any image at all could count as an answer
Visual	Does the answer have to be looked at rather than read?	The idea is fully delivered by the title, and the picture just illustrates it
Yours	Will you still care about this in four months, on a wet Tuesday?	You chose it because it sounded impressive

A question narrowed, step by step

Watch a real one get sharpened. A student starts with my grandmother. That is not a question; it is a person. It will produce one portrait and then stall.

1. My grandmother. Too broad. Nothing to change between pieces.
2. My grandmother's kitchen. Better: there is a place and a set of objects. But it is still a list, and a list produces one careful still life.
3. Can a set of objects stand in for a person who is not there any more? Now there is a question, and it can be answered badly. Some arrangements will read as a person; most will read as clutter. That gap is the content.
4. Pick the variable. Which objects, how many, lit how, cropped how tightly, drawn from life or from memory, in colour or not. Change one at a time.

That question generates a workable plan immediately: five drawings of single objects at life size; three of the same three objects rearranged; three drawn from memory a week apart, without looking; two photographed instead of drawn; two at wildly different scales. Fifteen pieces, one inquiry, a visible experiment in every one.

The same procedure, run on a question that starts too big

Another student wants to work on climate change. This is a real subject and a terrible inquiry, because at that scale every image becomes an illustration of a fact you already knew: a melting ice cap, a dying tree, a globe on fire. There is no variable, no possibility of being wrong, and nothing that could not be said better in a sentence.

Change one input, which is scale, and the whole thing flips. What does the tide line leave behind along one hundred metres of my nearest beach? Now the work is made of specific objects, gathered on specific days, in weather that changes. It can be catalogued, arranged, drawn at life size, printed, photographed at intervals across a term. It can fail: some collections will be visually dead. And it is still, unmistakably, about the same thing.

The move is always the same. Go down, not up. The smaller and more particular the question, the more work it produces.

Constraint as an engine

Two examples of how far a constraint can be pushed. Yves Klein spent much of his short career on a single ultramarine he had helped formulate, exhibiting canvases that were entirely and only that blue. Tate holds IKB 79, dated 1959. Bridget Riley, working from the early 1960s onward, built painting after painting out of nothing but repeated simple units, stripes, discs and curves, arranged so that a static surface appears to move. Tate holds a substantial group of her work.

Neither artist ran out of things to do. The constraint was not a limitation on the work; it was the apparatus that made the work possible, because once almost everything is fixed, tiny variations become legible. That is the same logic as Monet's window in the last lesson, applied harder.

Useful constraints to try for a term: one colour plus black and white; no line, only tone; only materials you did not buy; everything the same size; nothing larger than a postcard; no more than ten minutes per piece; only drawn with the hand you do not write with.

Twenty minutes, no editing

Here is the actual procedure. Set a timer for twenty minutes and do not stop or judge anything.

1. Write your name in the middle of a large sheet. Around it, write twenty nouns: objects, places and people that are genuinely present in your life. Not concepts.
2. Circle the five you would still be interested in looking at closely for an hour.
3. For each of the five, write three questions beginning with What happens if, Can a, or How much.
4. You now have fifteen candidate questions. Cross out any that fail the Specific test, which is usually about half.
5. Of the survivors, pick the two you would be slightly embarrassed to explain. Those are usually the real ones.

Then write a proposal. Three sentences, no more: what you are asking, what you will hold constant, what you will change. Under it, list your first five pieces in the order you will make them. That is your contract with yourself, and Lesson 17 will ask you to rewrite it as an artist statement once you know what actually happened.

In short: a good proposal is short, specific, and boring to read. It should describe an experiment, not advertise a feeling.

When to pivot, and how to do it honestly

Questions do drift, and a drift is not a failure. Around piece six or seven you will often notice that the interesting thing is not the thing you named. The rule is simple: you may change the question, but you must record the change and its date, and say what caused it.

What you must not do is silently redefine the inquiry so that everything you have already made appears to have been about the new thing all along. That is a lie told to a portfolio, and it is visible: the early work will not support the claim, and a reader can tell.

Common misconceptions

• "A serious question has to be about a serious subject." Scale of subject and quality of inquiry are unrelated. A question about the shadow under a chair can carry more work than a statement about war.
• "If I narrow the question I will run out of ideas." The reverse is true, and it is the most reliable finding in this course. Broad questions stall at two pieces; narrow ones generate twenty.
• "Constraints limit creativity." Klein worked one colour for years and Riley has worked with repeated simple units for six decades. Constraint is what makes small variation readable.
• "The proposal has to sound impressive." It has to be usable. Write it in the plainest language you can, because you will be reading it at eleven at night when you cannot think what to make next.
• "Changing the question means starting again." It means writing down when and why it changed. The earlier work stays, and it usually explains the shift.

The short version

Inquiries come from your own material: what you already look at, a constraint, a stubborn material, a place you can return to, a piece of evidence, or something that annoys you. Test each candidate for whether it is generative, specific, visual and genuinely yours. When a question is too broad, narrow it downward rather than reaching for a bigger subject, because particularity is what generates work. Lawrence knew his sixty panels before he mixed his first colour; Klein spent years on one blue; Riley has built six decades from repeated simple units. Write a three-sentence proposal naming the question, the constant and the variable, list your first five pieces, and keep it where you can see it. The next lesson turns to the book you will keep all of this in.

Sources

1. Tate. Yves Klein, IKB 79, 1959 [Paint on canvas]. Tate, London. tate.org.uk
2. Tate. Bridget Riley: artist page and works in the collection. Tate, London. tate.org.uk
3. College Board. AP Drawing: what the sustained investigation asks students to do. AP Students. apstudents.collegeboard.org
4. Lawrence, J. (1940-41). The Migration Series, sixty panels, tempera on hardboard. Museum of Modern Art, New York (even-numbered panels) and The Phillips Collection, Washington (odd-numbered panels).

When J. M. W. Turner died in 1851 he left the nation around three hundred sketchbooks, and they went to Tate as part of the Turner Bequest. Most of the pages in them are not beautiful. They are notes: a coastline in four lines, a wash of colour with a word written next to it, the same bridge from three positions on one sheet, numbers in the margin. Turner was not making small paintings in those books. He was collecting information he would need later, and the books survive because the information turned out to be worth having.

That is the correct model for your own sketchbook, and it is almost the opposite of what most students do with one.

The mistake almost everyone makes

The common failure is treating the sketchbook as a gallery: forty pages of careful, finished, medium-sized drawings, each one occupying a clean spread, nothing crossed out, nothing dated. It looks tidy. It is nearly useless, for two reasons.

First, if every page is finished, nothing in the book is thinking. There is no record of a decision because no decision is visible. Second, preciousness is expensive. If each page takes ninety minutes, you will make forty pages a year, and forty attempts is not enough attempts at anything.

Remember: the sketchbook is where work gets cheap. If a page is too valuable to ruin, it is in the wrong book.

Three jobs

Job	What goes in	How fast
Gathering	Observation from life, collected fragments, printed photographs at small size, colour notes, measurements, things picked up off the ground and taped in	Minutes per item
Testing	Thumbnails, material tests, the same composition tried six ways, one shape in five value ranges, an edge treated four different ways	Seconds to minutes
Recording	Dates on everything, one or two sentences on what changed and why, photographs of work in progress, the version you abandoned	Under a minute

A healthy book is mostly the middle row. If yours is mostly the first row, you are collecting without testing, which produces a lot of reference and no development.

Date everything, and here is why

Put the date on every page, in the same corner, every time. It takes three seconds and it converts a pile of drawings into a sequence.

Sequence is the whole point. Two similar drawings four weeks apart tell a reader something that the same two drawings undated cannot: that you returned to a problem and treated it differently the second time. Development is not a quality you can assert in a statement. It is a claim about time, and dates are the only evidence for it.

Annotation: opinion against decision

Most sketchbook writing is worthless because it is evaluative rather than operational. Compare:

Useless	Useful
I like this one.	The tilted horizon is doing the work here; the version with it level was inert.
This did not work.	The 2B will not go dark enough for the shadow side. Next attempt: 6B, and cut the background to three values so the head stays forward.
Experimenting with ink.	Ink on damp cartridge bleeds about 4mm. On the smooth paper it does not bleed at all and the line looks dead. Use the damp version for the outer marks only.
Inspired by Hokusai.	Took the low viewpoint from Hokusai: the horizon sits at one fifth from the bottom, so the sky does most of the picture.

The test for an annotation is whether it tells your future self what to do. "I like this" gives you nothing in March. "Cut the background to three values" is an instruction you can follow.

Keep it short. Two sentences per page is plenty, and a book crammed with paragraphs usually means the writing has become a substitute for the making.

Thumbnails, and why they must be small

A thumbnail is a compositional sketch about the size of a matchbox, two by three centimetres, made in under a minute. At that size you physically cannot draw detail, which is the point: you are forced to decide only where the large shapes go and how dark each area is.

The procedure before any substantial piece:

1. Draw six empty rectangles in the proportion of your final surface. Do not skip the proportion; a composition designed in a square will not transfer to a rectangle.
2. Fill each one in under sixty seconds, using no more than three values: light, mid, dark. No outlines, no detail.
3. Change one thing between each: the position of the main mass, the horizon height, the crop, the direction of the light, how much empty space you allow.
4. Turn the page upside down and look again. Composition problems are much more visible when you cannot read the subject.
5. Choose one. Write one sentence saying why, and what you gave up by not choosing the runner-up.

The upshot: six minutes of thumbnails routinely saves six hours of a drawing that was never going to work, because the problem was where the shapes sat and no amount of rendering fixes that.

What a working drawing looks like

Museums hold enormous numbers of drawings that were never meant to be seen, and they are worth looking at precisely because they are not finished. The Metropolitan Museum of Art holds a red chalk study by Annibale Carracci, Seated Youth Facing Right, Seen from the Back, with scribbles on the reverse. It is a study: a body worked out for use somewhere else. The lines are searching, there are several versions of the same contour, and nothing has been tidied.

That plurality of lines is not sloppiness. It is a record of the artist testing where an edge should be, and it is exactly what your own observational pages should look like. If your drawings have one confident line per edge from the start, you are probably drawing what you expect rather than what is there. Lesson 4 takes that apart.

A page structure that works

Use the spread, not the page. Left side: evidence. Right side: attempt.

• Left: the photograph you took, the object taped in, the measurements, the colour swatches, the note about the light.
• Right: six thumbnails, or the same shape tried four ways, with the chosen one circled.
• Across the bottom: the date, and one sentence saying what the next page will test.

That last sentence is the single most valuable habit in this lesson. Ending every session by writing what the next session will attempt removes the twenty minutes of drift that begins every studio hour.

Volume, and how to get it

Aim for six to ten pages a week during a portfolio year. That sounds impossible until you realise most of them take four minutes. A page of twenty thirty-second gesture drawings of people on a bus is a page. A page of eight thumbnails is a page. A page of one leaf drawn eight times at different speeds is a page.

If you are stuck, the reliable unstickers are: draw the same thing again smaller; draw it with the wrong hand; draw it in ten seconds; draw only the shadows; draw only the gaps between things; draw it from memory an hour later and then compare.

Digital sketchbooks

Working on a tablet is legitimate and this course does not disapprove of it. Two cautions. Digital work has no default date, so name and file everything with a date yourself, and keep versions rather than overwriting. And a digital canvas has infinite undo, which quietly removes the record of your mistakes; export a flattened copy of the failed state before you fix it, or the evidence disappears.

Whatever you use, photograph your physical pages regularly. Sketchbooks get lost, rained on and left on buses, and a portfolio built on a book that no longer exists is a hard conversation.

Common misconceptions

• "The sketchbook should look impressive." It should look used. A reader is looking for change over time, not for forty small finished drawings.
• "Thumbnails are for people who cannot draw well yet." They are a compositional tool, and skipping them is why competent drawings fail at the level of arrangement.
• "Annotation means writing what you think of the piece." It means writing what you will do next. Opinions age badly; instructions do not.
• "Undated pages are fine because I remember the order." You will not, and more importantly a reader cannot check. Development is a claim about time.
• "Working digitally means I do not need to keep failures." Infinite undo erases exactly the evidence that shows your thinking. Export the broken version first.

What to carry forward

A sketchbook has three jobs: gathering material, testing possibilities and recording decisions, and a healthy one is mostly testing. Date every page, because sequence is the only evidence of development. Write annotations that instruct rather than evaluate, keep them to a sentence or two, and end each session by naming what the next one will attempt. Make thumbnails small enough that detail is impossible, six at a time, in the proportion of the final surface, changing one variable between them. Expect your observational pages to carry several searching lines per edge, as Carracci's studies do, rather than one confident line placed from memory. Aim for volume over polish, and photograph everything. Module 2 begins the drawing itself, starting with the hardest problem in observation: seeing what is actually in front of you.

Sources

1. Tate. Joseph Mallord William Turner: artist page and the Turner Bequest. Tate, London. tate.org.uk
2. Carracci, A. Seated Youth Facing Right, Seen from the Back [Red chalk with white chalk highlights]. The Metropolitan Museum of Art, New York. metmuseum.org
3. Khan Academy. Art history basics: looking at and describing works of art. Khan Academy. khanacademy.org
4. Getty. Publications and reports: freely available books on drawings and drawing practice. J. Paul Getty Trust. getty.edu

Put a mug on the table in front of you and lower your head until your eye is level with the rim. The opening is not a circle. It is a flat ellipse, perhaps three millimetres tall, almost a line. Now raise your head twenty centimetres and it opens into a fat oval. Almost everyone, asked to draw that mug, draws something rounder than what is actually there, because they are not drawing the mug on the table. They are drawing the mug they know about.

This is the central problem of observational drawing, and it is not a problem of hand skill. Your hand is fine. Your eye is fine. What gets in the way is a stored symbol, built up since childhood, that fires before you have finished looking.

Why symbols win

By about ten years old most people have a compact set of drawing schemas: an eye as a pointed almond with a circle in it, a hand as a palm with five sausages, a house as a square with a triangle, a table with four legs even when only three are visible from where you sit. These schemas are efficient. They let you communicate an idea instantly, which is what they were built for.

They are also wrong in a specific way: they encode what a thing is, not what it looks like from here. When you sit down to draw from life, the schema arrives first, and unless you actively block it, you will draw the schema and then adjust it slightly towards the subject. That is why beginner drawings often look confident and inaccurate at the same time.

Key idea: observational drawing is not learning to move a pencil better. It is learning to delay the naming of what you see for long enough to record its shape.

Contour drawing, and the three versions of it

A contour is the edge where one form ends and another begins. Drawing contours slowly is the standard method for getting past symbols, and it comes in three forms.

Type	What you do	What it trains
Blind contour	Eyes on the subject only. Never look at the page. Move the pencil at exactly the speed your eye travels along the edge, without lifting it.	The link between eye and hand. The result is always distorted and that is not the point.
Modified contour	The same slow tracking, but you may glance at the page at chosen moments to re-register your position.	Accuracy, while keeping the slow looking. This is the one you will actually use.
Cross-contour	Lines that travel across the form rather than around it, as though a fly were walking over the surface.	Understanding volume. It forces you to think about where the surface turns.

Blind contour feels ridiculous for about four minutes and then something changes: you start noticing that edges wobble, that the line between a thumb and its background is not smooth, that the gap between two fingers has a shape. Fifteen minutes of it before a serious drawing measurably improves what follows.

The instruction that matters most is the speed one. Move the pencil as slowly as your eye travels. If the pencil arrives somewhere before your eye does, the hand is running on the schema again.

What a contour can carry on its own

Line alone, with almost no shading, can describe a whole figure. The Metropolitan Museum of Art holds a graphite portrait by Ingres, Madame Félix Gallois, dated 1852, in which the drawing of the head and hands is carried almost entirely by an edge of continuously varying weight, with the rest of the sheet left comparatively bare. Nothing is filled in for its own sake. Where the line thickens, the form turns away; where it thins, the light runs along it.

That is the ambition: a line that reports information rather than outlining a shape. A uniform line of constant thickness is a border. A line that changes weight is a description.

Negative space, and why it defeats the symbol

Negative space is the shape of the gap: the area between a chair's legs, the wedge of wall between an arm and a body, the shape of the sky enclosed by two branches.

Drawing the gaps works for a very simple reason. Your brain has no stored symbol for the shape between a chair leg and a stretcher. It has never needed one. So when you draw that shape, you are forced to look at it, and getting the gaps right automatically puts the solid parts in the right place.

The procedure: look at your subject and mentally darken everything that is not the object. Draw only those shapes, edge to edge, as though they were the subject. Do not draw the chair at all. When the gaps are complete, the chair is already there.

This is also the fastest way to fix a drawing that feels wrong but you cannot say why. Look at one negative shape in your drawing and the same one in life, and compare them directly. The error will usually be obvious within seconds, because gaps are easy to compare and objects are not.

What matters here: you cannot check a drawing against a thing you have named. You can check it against a shape you have not.

The upside-down trick

Betty Edwards, in Drawing on the Right Side of the Brain (1979), popularised an exercise that makes this concrete: take a line drawing, turn it upside down, and copy it upside down. Almost everyone produces a noticeably better copy than they would the right way up.

The explanation offered in the book, about left and right brain hemispheres, has not held up well as neuroscience. The exercise still works, and the reason is the one above: upside down, the marks stop being a nose and a chin and become shapes and angles, and shapes and angles are what you can actually copy. You can apply this at any time by turning your own drawing upside down to check it, which is the same thing you did with thumbnails in the last lesson.

Ellipses, in two rules

Round things in perspective are where symbol errors are most visible, and two rules cover almost every case.

1. An ellipse gets rounder the further it is from your eye level. A glass at eye level shows a nearly flat rim. Look down into it from above and the rim approaches a circle. A stack of cups therefore shows a different ellipse at every height, and drawing them all the same is an instant giveaway.
2. The minor axis rule. Every ellipse has a long axis and a short one, and the short axis lies along the axis of the object. For an upright glass the short axis is vertical, so the ellipse is symmetrical left to right and its widest point sits halfway up. Drawing a lopsided ellipse with one corner pointed is the commonest object-drawing error there is.

Ellipses should have no corners at all. If your line has a point where it changes direction sharply, you have drawn two arcs meeting, not an ellipse. Practise them fast and lightly, going round several times, rather than slowly in one careful pass.

Edges are not all the same

Once you are looking properly, you will notice that edges vary enormously, and treating them all identically is what makes a drawing look cut out from paper.

Edge	Where it happens	How to draw it
Hard	A sharp form against a contrasting background; the edge of a book	Crisp, definite, often the darkest line in the area
Soft	A form turning gradually away, as on a cheek or a sphere	Blended, no line at all, tone meeting tone
Lost	Where a dark form meets a dark background and the boundary genuinely disappears	Leave it out. Do not draw what you cannot see.
Found	Where a lost edge re-emerges into light	Let it appear gradually, so the eye completes it

Lost edges are the hardest thing to persuade a beginner to do, because leaving out an edge you know is there feels like an error. It is the opposite: drawing an edge you cannot see is the error, and it is the reason so many careful drawings look stiff.

Common misconceptions

• "I cannot draw because my hand is not steady enough." Almost never the problem. The problem is that the symbol arrives before the looking finishes.
• "Blind contour is a warm-up game." It is a corrective exercise with a measurable effect on the drawing that follows. It is meant to look wrong.
• "Negative space is an abstract idea for advanced students." It is the most practical checking tool in the course, because the gaps carry no names and therefore no symbols.
• "An outline should be an even line." An even line is a border. A line that varies in weight reports where the form turns and where the light falls.
• "If I know the edge is there I should draw it." Only if you can see it. Lost edges are how a drawing sits in space instead of on top of the paper.

Pulling it together

The obstacle in observational drawing is a stored symbol that fires before you have finished looking, which is why beginner drawings are often confident and wrong at the same time. Contour drawing slows the eye down: blind contour rebuilds the eye-hand link, modified contour keeps that slowness while allowing accuracy, and cross-contour makes you think about where a surface turns. A line that varies in weight, as in the Ingres portrait at the Metropolitan Museum, describes; an even line merely encloses. Negative space defeats the symbol because gaps have no names, which also makes it the fastest way to diagnose a drawing that feels wrong. Ellipses get rounder away from eye level, sit symmetrically about their short axis, and never have corners. And edges come in four kinds, of which the one to practise hardest is the one you leave out. Next lesson replaces guessing at proportion with measuring it.

Sources

1. Wikipedia contributors. (2024). Contour drawing. Wikipedia. en.wikipedia.org
2. Wikipedia contributors. (2024). Negative space. Wikipedia. en.wikipedia.org
3. Ingres, J. A. D. (1852). Madame Félix Gallois [Graphite on buff wove paper]. The Metropolitan Museum of Art, New York. metmuseum.org
4. Edwards, B. (1979). Drawing on the Right Side of the Brain. Los Angeles: J. P. Tarcher.

Albrecht Dürer published a treatise on measurement in Nuremberg in 1525, and one of its woodcuts shows two men drawing a lute. Between them stands a frame with a grid of threads; one man moves a pointer over the instrument while the other plots each point onto paper. The Metropolitan Museum of Art holds an impression of it, catalogued as The Draughtsman of the Lute. The machine looks absurd. It is also honest about something artists have known for five hundred years: getting proportion right is a measuring problem, and the eye alone is not reliable at it.

Why proportion is the error people notice

A viewer who cannot say anything technical about a drawing will still know instantly that a portrait is wrong. Almost always the fault is proportional: the head too narrow, the eyes too high, the arm too short, the table too shallow. Rendering cannot repair it. You can shade a badly proportioned drawing for six hours and it will remain a badly proportioned drawing, beautifully shaded.

So measure first, render later. In practice that means the first twenty minutes of any observational drawing are spent placing a small number of large relationships, with a very light pencil, and checking them before anything is committed.

The point: accuracy is decided in the first ten per cent of the drawing. Everything after that is either building on a correct structure or decorating a wrong one.

The pencil at arm's length, done properly

The technique everyone half-knows, with the details that make it actually work:

1. Lock your arm straight. Elbow completely extended, every single time. If your arm bends by five centimetres between measurements, all your measurements are on different scales and are worthless.
2. Close one eye. Two eyes give you two slightly different positions. Use the same eye all session.
3. Hold the pencil perpendicular to your line of sight, flat on to you, not angled away, or it will read shorter than it is.
4. Align the tip with one end of what you are measuring, slide your thumb down the shaft until the nail sits at the other end. That distance is now your measurement.
5. Do not move your thumb while you turn the pencil to compare it with something else.
6. Stand or sit in exactly the same spot. Mark the floor with tape. Twenty centimetres of drift changes every relationship in the subject.

Two things this does not give you. It does not give you the true size of anything, only ratios; and it does not work if you are closer to some parts of the subject than the method assumes, which is why a very large subject seen from very close is hard to measure this way.

Choose a unit and stay with it

Comparative measurement means picking one dimension in the subject as your unit and describing everything else in multiples of it. For a figure, the head. For a still life, the width of the largest jar. For a room, the height of a door.

Then the questions you ask become answerable: how many heads tall is this figure? How many jar-widths from the left edge to the bowl? Is this shape wider than it is tall, or taller than it is wide, and by how much? That last question, asked before you draw anything, prevents most beginner errors on its own.

Standard relationship	Rough canon	Warning
Adult figure height	About seven and a half heads	Idealised figures in art are often drawn at eight; real people vary a lot
Eyes on the head	Roughly halfway down the whole head, top of skull to chin	Nearly everyone places them too high, because hair is discounted
Ear	Spans from brow line to base of nose	True in a straight-on view only
Hand	About the length of the face, chin to hairline	Check it on yourself before you believe it
Width of shoulders	Roughly two heads for an adult	Varies with build more than the others

These are starting expectations, not rules. Their real value is diagnostic: when your drawing disagrees with the canon by a lot, look again, because you have probably made a symbol error rather than an observation.

Angles: the clock face and the plumb line

Most people can judge whether a line is roughly vertical. Almost nobody can judge a thirty-degree slope by eye. Two methods fix this.

The clock face. Hold your pencil up at arm's length and rotate it until it lies along the edge you are looking at. Now, without changing that angle, notice what time it would read on a clock: half past four, twenty past one. Bring the pencil down to the page keeping the angle locked, and lay it against the paper to transfer it. Comparing the angle to something fixed, vertical or horizontal, is far more accurate than trying to remember a slope.

The plumb line. Hold the pencil vertically, or use an actual weighted string, and let it pass through a chosen point in your subject. Now look at what else it crosses. In a standing figure, drop a plumb line from the pit of the neck: does it fall through the ankle, in front of it, behind it? That single observation fixes the whole balance of a pose. Do the same horizontally: from the corner of a table, what else in the room is at the same height?

Plumb lines are also how you check a drawing you have already made. Hold the pencil vertically against your paper through a point, then hold it vertically in the same place in the subject, and compare what lines up. Disagreements are errors, and they are specific enough to fix.

Sight-size, and the viewfinder

Sight-size is a stricter method: you place your drawing beside the subject and set the drawing so that the image on it is the same size as the subject appears from a fixed viewing spot. Then you step back to that spot, and you can compare drawing and subject directly, in the same glance, at the same scale. It removes all the mental scaling that ordinary comparative measurement demands, and it makes errors glaringly obvious. Its cost is rigidity: the spot, the easel and the subject cannot move, and the drawing can only be the size the setup dictates.

A viewfinder is a much simpler tool and worth making. Cut a rectangular hole in a piece of card, in the same proportion as your paper, and hold it up at a fixed distance from your eye. It crops the world for you, so you can decide the composition before you draw a line, and its edges give you two reliable straight references to judge angles against.

Debug: the head that is correct and still wrong

A student draws a portrait. The eyes are well observed, the nose is right, the mouth is convincing, the shading is careful. The drawing still reads as a slightly alien child. What went wrong?

Measure the head. In nearly every case of this failure, the distance from the eye line to the top of the skull is too short. The reason is attention: the features are interesting, so they get looked at and measured, while the cranium is a large plain dome that the eye skips over. The symbol for a head puts the eyes high because that is where a cartoon puts them.

The fix is a measurement, not more looking: check the eye line against the halfway point of the whole head including hair. Then check the same thing on any portrait in a museum. The Metropolitan Museum's Vermeer, Young Woman with a Water Pitcher, of about 1662, is a useful test case, because the headdress makes the top of the skull unusually easy to locate.

The same error runs the other way with the lower face: beginners often make the distance from nose to chin too long. Measure both, every time, until you stop making the error without checking.

A note on grids and tracing

Transferring an image with a grid, or tracing a projection, is a legitimate technique with a long history; Dürer's threaded frame is its ancestor and artists have used camera obscuras, projectors and grids for centuries. There is no moral problem with it.

There is a learning problem. A grid lets you copy accurately without ever having to judge a relationship, so it builds no measuring skill at all. Use it for transferring a composition you have already designed, at a different scale, when you are working from your own reference. Do not use it as a way of avoiding observational drawing, because the skill it substitutes for is the skill this module exists to build.

Bottom line: measure with the pencil to learn to see; use a grid to move an image you already understand.

Common misconceptions

• "Measuring is a crutch for people who cannot draw." Every representational tradition has used measuring devices, and the ones who did not still measured with a stick.
• "The pencil measurement gives you the real size." It gives ratios only, and only if your arm is locked and you do not move.
• "Getting the features right will make the portrait right." The commonest portrait error is in the cranium, which nobody looks at because it is featureless.
• "The eight-heads figure is the correct proportion for people." Eight heads is an idealised canon. Real adults are closer to seven and a half, and vary.
• "You can fix proportions later with shading." You cannot. A badly proportioned drawing shaded well is still badly proportioned.

Recap

Proportion is the error viewers actually see, and rendering never repairs it, so measure in the first minutes and commit nothing until the large relationships check out. The pencil method works only if the arm is locked straight, one eye is closed, the pencil is flat on to your view, and you do not move from your spot. Choose one unit, express everything in multiples of it, and use the standard canon relationships diagnostically rather than as rules. Judge angles against vertical and horizontal using the clock-face transfer, and use plumb lines both to read the subject and to check your own drawing. Sight-size removes mental scaling at the cost of rigidity; a card viewfinder is cheap and does much of the same job. When a portrait reads as wrong despite good features, measure the skull. And treat grids as a transfer tool, not as a way of learning to see. Next: turning a flat, accurate outline into something that has a light on it.

Sources

1. Dürer, A. The Draughtsman of the Lute [Woodcut]. The Metropolitan Museum of Art, New York. metmuseum.org
2. Wikipedia contributors. (2024). Body proportions. Wikipedia. en.wikipedia.org
3. Tate. Foreshortening. Tate Art Terms, London. tate.org.uk
4. Vermeer, J. (ca. 1662). Young Woman with a Water Pitcher [Oil on canvas]. The Metropolitan Museum of Art, New York. metmuseum.org
5. Dürer, A. (1525). Underweysung der Messung mit dem Zirckel und Richtscheyt. Nuremberg.

In Georges de La Tour's The Penitent Magdalen, of about 1640, at the Metropolitan Museum of Art, there is exactly one light source and it is inside the picture: a candle flame, its base concealed behind a skull. Everything else in the painting is a report on how far each surface is from that flame and at what angle it faces it. The cheek is bright because it turns towards the candle. The far wall is nearly black because it does not. Take the candle away and the painting has no subject at all.

That is what value does. Line tells you where things are. Value tells you what shape they are, where the light is, and how far away everything sits.

Value is not colour

Value is how light or dark something is, independent of what colour it is. A mid-yellow and a mid-blue can have identical value, which is why a photograph of them in black and white shows two identical greys. Learning to see value separately from colour is the single hardest perceptual skill in this module, because your eye is extremely interested in hue and only moderately interested in lightness.

The tool is a value scale. Make one now and keep it: a strip of paper divided into nine boxes, box 1 left as white paper, box 9 as dark as your material will go, and the seven between them stepping evenly. Evenly is the difficult part; most people compress the middle and stretch the ends. Hold the finished strip against anything and you can read its value by matching.

Why this matters: without a physical reference, you will judge every value relative to the one next to it and drift steadily lighter across a drawing. With a strip in your hand, you are matching against a fixed standard.

The six zones on a lit form

Set an egg or a plain white ball on a table with a single lamp to one side, and turn the other lights off. Every zone below will be visible, and they always occur in this order going away from the light.

Zone	Where it is	Typical value	Note
Highlight	Where the surface faces the light most directly	1, or the white of the paper	Small, and sharp on shiny things, broad and soft on matt ones
Light	The lit plane around the highlight	2 to 3	Most of the light side
Halftone	Where the form begins to turn away	4 to 5	The zone that does most of the describing
Core shadow	The band where the surface turns fully away, also called the terminator	7 to 8	Usually the darkest part of the object itself
Reflected light	Beyond the core shadow, light bouncing back off the table	6	Never as light as the light side. This is the rule beginners break.
Cast shadow	The shadow the object throws onto the surface	8 to 9 near the object, lighter further away	Hardest edge where it meets the object, softening with distance

Two consequences follow immediately, and both are counterintuitive.

First, the darkest part of the object is not at its edge. It is the core shadow, which sits inside the silhouette with reflected light beyond it. Drawing an object as dark all the way to the contour makes it read as flat and cut out. The strip of reflected light along the shadow edge is what makes a sphere look round.

Second, reflected light must stay dark. It belongs to the shadow family. If you make it as light as the halftone, the form flattens instantly and the object looks like it is glowing from within. Keep every value in shadow darker than every value in light, without exception. That single rule fixes more student drawings than any other.

Cast shadows have shape and edges

A cast shadow is not a grey puddle. It is the silhouette of the object projected onto whatever surface receives it, distorted by the angle of that surface, and it obeys rules:

• Its edge is sharpest where it touches the object and gets progressively softer the further it travels.
• It is darkest immediately under the object, where almost no ambient light can reach. That very dark contact point is sometimes called the occlusion shadow, and putting it in is what makes an object sit on a table rather than float above it.
• A small hard light source gives hard-edged cast shadows; a large diffuse source, such as an overcast sky or a window, gives soft ones.
• It follows the form it lands on, so it bends when it crosses an edge and stretches when it crosses a receding plane.

Local value against value structure

Now the idea that separates people who can draw a lit scene from people who can only draw objects.

Local value is how light or dark a thing is in itself: white paper is light, a black mug is dark. Value structure is how light or dark it appears in the actual light falling on it now. These are different, and when they conflict, value structure wins.

Object	Local value	In this light	What you must draw
White cup, in shadow	1	Reads as about 6	6
Black mug, in direct light	9	Reads as about 5	5
White wall, far from the window	1	Reads as 5 or 6	5 or 6

So a white object in shadow can be genuinely darker than a black object in light. Beginners refuse to believe this and draw what they know: white things pale, dark things dark. The result is a drawing with no light in it, because light is precisely the thing that overrides local value.

The way to see it is to squint hard until your eyelashes blur the scene into a handful of masses. Squinting throws away detail and colour and leaves you with value groups. Do it constantly.

The core of it: draw the light, not the objects. If the light is described correctly, the objects appear on their own.

Three values before nine

A reliable working method, and the fastest route to a drawing that reads from across the room:

1. Squint until the scene reduces to three groups: light, mid, dark. Ignore everything else.
2. Block those three in flat, with hard boundaries and no blending. The drawing will look crude. Leave it crude.
3. Check the three-value version by turning it upside down or photographing it. If it does not read at this stage, no amount of refinement will save it.
4. Only now subdivide: split the mid group into halftones, find the core shadow inside the dark group, put in the reflected light, and lift the highlight with an eraser.

Working the other way, detail first, produces drawings that are full of careful information and have no structure, and there is no way to add structure afterwards without destroying the detail.

Two historical strategies worth knowing

Chiaroscuro is the general use of strong light and dark contrast to model form. Its extreme version, in which most of the picture sits in deep shadow and a narrow light picks out what matters, is often called tenebrism, and Caravaggio is its most famous practitioner. The Metropolitan Museum holds The Denial of Saint Peter, dated 1610, in which three heads and a pair of hands are lit and everything else is given up to darkness. Note what that does compositionally: value alone decides where you look.

Sfumato is the opposite instinct: transitions so gradual that no edge is visible anywhere, associated above all with Leonardo. Where Caravaggio separates light and dark with a knife, sfumato dissolves the boundary. Both are complete answers to the same problem, which is a useful thing to notice: there is no single correct way to handle an edge.

Materials for value

Material	Range	Good for
Graphite, H to 2B	Values 1 to 5 only	Light structure, halftones, precise work
Graphite, 4B to 8B	Reaches 7 or 8, and shines	Darks, though the sheen can misread as light
Vine or willow charcoal	Full range, very erasable	Blocking in three values fast; wipes back with a cloth
Compressed charcoal	Reaches a true 9, hard to remove	The darkest darks, laid in last
Kneaded eraser	Subtracts	Lifting highlights; it is a drawing tool, not a correction tool

If your drawing looks weak, the usual cause is that its darkest value is only a 6. Most beginners never use the bottom third of the scale. Put a true dark somewhere and the whole range wakes up.

Common misconceptions

• "Shadows are grey." Shadows are a value relationship, and often they carry strong colour. What matters is that they stay darker than everything on the light side.
• "Reflected light is bright." It is the lightest part of the shadow, which is not the same thing. Make it too light and the form goes flat.
• "The edge of an object is its darkest point." The core shadow sits inside the silhouette, with reflected light between it and the contour.
• "A white object should be drawn light." Only if the light is on it. A white cup in shadow is darker than a black mug in sunlight, and drawing otherwise removes the light from the picture.
• "You should outline first and shade inside the lines." That produces a coloured-in shape. Block three values first and let the edges emerge from where the values meet.

Putting it together

Value is lightness independent of hue, and a nine-step scale made by your own hand is the reference that stops you drifting. A single form under a single light shows six zones in a fixed order: highlight, light, halftone, core shadow, reflected light and cast shadow, with the darkest part of the object inside the silhouette rather than at its edge, and every shadow value darker than every light value. Cast shadows are sharpest and darkest where they touch the object. Local value tells you what a thing is; value structure tells you what it looks like under this light, and value structure always wins, which is why a white cup in shadow can be darker than a black mug in sunlight. Squint, block three values flat, check that the drawing reads, and only then subdivide. Caravaggio and Leonardo represent the two extremes of edge handling, and both are correct. Module 3 turns from the light on a form to the space that form sits in.

Sources

1. de La Tour, G. (ca. 1640). The Penitent Magdalen [Oil on canvas]. The Metropolitan Museum of Art, New York. metmuseum.org
2. Caravaggio (Michelangelo Merisi). (1610). The Denial of Saint Peter [Oil on canvas]. The Metropolitan Museum of Art, New York. metmuseum.org
3. Tate. Chiaroscuro. Tate Art Terms, London. tate.org.uk
4. Wikipedia contributors. (2024). Sfumato. Wikipedia. en.wikipedia.org
5. Wikipedia contributors. (2024). Lightness. Wikipedia. en.wikipedia.org

Some time around 1413, Filippo Brunelleschi stood in the doorway of Florence Cathedral and painted the Baptistery opposite onto a small panel. Then he drilled a hole through the panel at the point where the lines converged. A viewer was told to stand exactly where he had stood, hold the panel with the painted face turned away, look through the hole from behind, and hold up a mirror at arm's length. The mirror showed the painting. Drop the mirror and you saw the building itself. The demonstration existed to prove that the two matched.

Twenty-two years later Leon Battista Alberti wrote the method down in De pictura, and European painting spent the next four centuries working out what to do with it.

The horizon is your eye level, and nothing else

Everything in this lesson rests on one fact that students consistently get wrong: the horizon line is at the height of your eyes. Not the middle of the page, not where the ground meets the sky in the scene, not where the far wall is. If you sit on the floor, the horizon is low and you will see the tops of very few things. If you stand on a chair, it is high and you look down onto tables.

Two consequences that let you check any drawing instantly:

• Anything above your eye level shows you its underside, and its receding edges slope down towards the horizon.
• Anything below your eye level shows you its top, and its receding edges slope up towards the horizon.
• Anything exactly at your eye level shows as a flat line, which is why a tabletop disappears to nothing when you crouch to its height.

Key idea: perspective is not a set of rules about buildings. It is a description of what happens when parallel lines are seen from a particular position, and your eye level is that position.

One point, two points, three points

System	When it applies	What converges	Typical subject
One-point	One face of the object is parallel to the picture plane, square on to you	Only the edges running away from you, all to a single vanishing point	A corridor, a road, a room seen face on, railway tracks
Two-point	The object is turned, so no face is square on, but it is upright	Two sets of horizontal edges, to two vanishing points on the horizon. Verticals stay vertical.	A building seen from a corner, a box on a table, a street junction
Three-point	You are looking sharply up or down	The two horizontal sets, plus verticals converging to a third point above or below	Looking up at a tower, looking down a stairwell

Almost everything you draw this year is one-point or two-point. The most useful thing to internalise about two-point is that verticals do not converge. If you find yourself tilting the vertical edges of a box towards each other, you have accidentally started drawing in three-point, and unless you are looking distinctly up or down, it will read as wrong.

Building a box in two-point, step by step

1. Draw the horizon line where your eyes are. Place two vanishing points on it, well apart, ideally off the edge of the paper. Vanishing points that are close together produce violent distortion, which is the commonest amateur perspective error.
2. Draw the vertical edge that is nearest to you. Its length sets the scale of everything.
3. From the top and bottom of that vertical, draw lines to the left vanishing point and to the right vanishing point. You now have four lines and the box's near corner.
4. Decide how deep the box goes and put in two more verticals, one on each side, cutting those lines.
5. From the tops of those two new verticals, draw back to the opposite vanishing points. Where they cross is the far top corner, and the box is closed.

Everything else is this box repeated: a table is a box, a room is a box you are inside, a head can be constructed inside a box, a street is a row of them.

Dividing a receding plane

Here is the trick that makes perspective genuinely useful rather than merely correct. You often need to divide a receding surface evenly: the panes of a window, the boards of a floor, the fence posts along a road. Even divisions in reality are not even in the drawing, because the far ones are smaller. Two methods:

To find the centre of any receding rectangle, draw both diagonals. They cross at the true centre in perspective, and a line through that point to the vanishing point divides the rectangle in half. This works at any depth and needs no measurement at all.

To repeat equal spacings, take a rectangle you already have, find its centre with the diagonals, draw a line from the top corner of the near side through that centre point, and continue it until it hits the receding line. That is where the next equal division falls. Repeat to march a fence into the distance with correctly shrinking gaps.

Remember: the diagonals of any rectangle cross at its centre, and this remains true when the rectangle is in perspective. It is the single most useful construction in this lesson.

Aerial perspective: the other way to make depth

Linear perspective handles the geometry. It does nothing at all about atmosphere, and a drawing built on lines alone often still reads flat. Aerial or atmospheric perspective is what air does to distance, and it is four separate effects:

With distance	What happens	How to use it
Contrast	Reduces; distant darks are not very dark	Keep your full value range for the foreground only
Edges	Soften	Sharpest edges nearest the viewer
Value	Compresses towards the value of the sky	Distant hills sit in a narrow band of middle values
Colour	Cools and desaturates, drifting towards blue	Warm and saturated in front, cool and grey behind

These effects are strong enough to override geometry. A drawing with no vanishing points at all can read as deep space if the near things are dark, sharp and saturated and the far things are pale, soft and grey. Combining the two systems is what makes a landscape convincing.

Two paintings, two different solutions

Paolo Uccello's The Battle of San Romano, in the National Gallery in London, is one of the most conspicuous displays of perspective in fifteenth-century painting. Broken lances lie on the ground arranged in a rough grid running back into the picture, and fallen figures are foreshortened. It looks less like a battle than like a demonstration of a new technology, which in part it is: Uccello was, by contemporary report, obsessive about perspective.

Compare Jan van Eyck's The Arnolfini Portrait of 1434, also in the National Gallery. The room recedes convincingly and the convex mirror on the back wall shows the whole space including two figures in the doorway. But the receding lines do not all meet at a single vanishing point. Van Eyck was working empirically, by eye and by extraordinary observation, rather than by Alberti's construction. The room is completely convincing anyway.

That comparison is worth holding onto. Geometric correctness and visual conviction are related but not identical, and one painted twenty years before the other is not therefore worse.

What linear perspective is not

It is a model of one situation: a single motionless eye, one fixed viewpoint, a flat picture plane. Real vision involves two eyes that move constantly, and the periphery of your visual field is not projected onto a flat plane at all. That is why a photograph taken with a very wide lens looks distorted at the edges even though it is geometrically correct.

And it is a convention, not a universal law of pictures. Chinese handscroll landscape painting uses shifting viewpoints so the viewer travels through the scene as the scroll unrolls. Egyptian tomb painting shows each part of a body from its most legible angle. Medieval European painting scales figures by importance rather than by distance. Cubism deliberately holds several viewpoints at once. None of these is a failed attempt at perspective; they are different decisions about what a picture is for.

So: learn the construction properly, because you cannot break a rule you cannot execute. Then decide, deliberately, whether your investigation wants it.

Common misconceptions

• "The vanishing point goes in the middle of the picture." It goes wherever you are looking, on the horizon line, and putting it off-centre is usually a better composition.
• "The horizon is where the ground meets the sky." It is your eye level, which may be well above or below any visible landscape edge, and exists indoors too.
• "Verticals converge in two-point perspective." They do not. Converging verticals means three-point, which only applies when you are looking sharply up or down.
• "Vanishing points should be inside the paper." Close vanishing points cause severe distortion. Put them off the edge and use a long ruler or a piece of string.
• "Perspective equals realism." It is one convention among several, and empirically constructed spaces such as van Eyck's can be more convincing than geometrically exact ones.

What you now know

Linear perspective describes what happens to parallel lines seen from one fixed position, and the horizon line is that position: your eye level. Things above it slope down to the horizon, things below it slope up, things level with it flatten to a line. One-point applies when a face is square on to you, two-point when the object is turned but upright, and three-point only when you look sharply up or down; in two-point, verticals stay vertical and vanishing points belong far apart. The diagonals of a rectangle cross at its true centre even in perspective, which lets you divide any receding plane without measuring. Aerial perspective adds the four atmospheric effects, reduced contrast, softer edges, compressed values and cooler colour, and it can carry depth on its own. Brunelleschi proved the geometry around 1413 and Alberti wrote it down in 1435, but van Eyck built a convincing room without it, and several major traditions chose not to use it at all. Next: the other half of what makes a picture, which is colour.

Sources

1. The National Gallery. Paolo Uccello, The Battle of San Romano, about 1438-40. The National Gallery, London. nationalgallery.org.uk
2. The National Gallery. Jan van Eyck, The Arnolfini Portrait, 1434. The National Gallery, London. nationalgallery.org.uk
3. Wikipedia contributors. (2024). Perspective (graphical). Wikipedia. en.wikipedia.org
4. Wikipedia contributors. (2024). Filippo Brunelleschi. Wikipedia. en.wikipedia.org
5. Alberti, L. B. (1435). De pictura. Florence.

In 1824 the chemist Michel Eugène Chevreul was put in charge of dyeing at the Gobelins tapestry works in Paris, and he immediately started receiving complaints about the blacks. Weavers said certain black yarns looked weak. Chevreul tested the dye and found nothing wrong with it. The blacks were only weak when they sat next to particular colours: what the weavers were seeing was not a property of the yarn at all but an effect of its neighbours. He published the finding in 1839 as the law of simultaneous contrast of colours, and it went on to shape Impressionist and Neo-Impressionist painting.

Start here, because it is the most useful fact in colour theory: no colour has a fixed appearance. Every colour is altered by what is next to it.

Three dimensions, not one

Beginners think about colour in one dimension, hue, and that is why their colour looks arbitrary. Every colour has three independent properties.

Dimension	What it means	Question to ask
Hue	Which colour it is: red, orange, blue-green	Where does it sit on the wheel?
Value	How light or dark it is	What number is it on your nine-step scale?
Saturation (chroma)	How pure or how grey it is	How far is it from a neutral grey of the same value?

Two paints can share a hue and look nothing alike because one is pale and grey and the other is dark and intense. When a colour in your work is not behaving, name which of the three dimensions is wrong before you touch the paint. Nine times out of ten it is value, and nine times out of ten students try to fix it by changing hue.

The point: value carries the structure, saturation carries the emphasis, hue carries the mood. Get them in that order.

Which colour wheel, and why there is more than one

Model	Primaries	Used for	Honest status
RYB	Red, yellow, blue	Traditional artists' wheel, paint mixing at school level	A useful convention, not physics; it cannot reach some colours
CMY	Cyan, magenta, yellow	Printing, and serious paint mixing	Subtractive, and reaches a noticeably wider range than RYB
RGB	Red, green, blue	Screens, digital work, light	Additive: the primaries add up to white, not to black

The difference between the last two is worth understanding, because you will move between paint and screen constantly. Additive mixing is what light does: shine red, green and blue light on the same spot and you get white. Subtractive mixing is what pigment does: each pigment absorbs part of the spectrum, so mixing them removes more and more light and heads towards black, or in practice towards a muddy brown.

This is why a colour that glows on your screen often cannot be mixed in paint at all. The screen is emitting light; the paper is only reflecting what falls on it.

Complementaries: the pair that does two opposite jobs

Complementary colours sit opposite each other on the wheel. On the traditional artists' wheel the pairs are red and green, blue and orange, yellow and violet.

They do two things, and they are the opposite of each other:

1. Side by side, they maximise contrast. Each makes the other look more intense. A small patch of orange in a blue field is the loudest thing you can do with colour, which is why it turns up whenever a painter wants your eye to go somewhere specific.
2. Mixed together, they neutralise. Adding a little of a colour's complement drains its saturation without changing its value much, which is how you grey something down without adding black. Adding black darkens and deadens; adding the complement greys while keeping the colour alive.

That second use is the professional one. If your greens are shrieking, do not add black or white. Add a touch of red. Most of the subtle colour in serious painting is made this way: a hue pulled part of the way towards its opposite.

Simultaneous contrast, in practice

Chevreul's law, restated for a studio: a colour shifts in appearance according to its surroundings, and it shifts away from its neighbour.

• A neutral grey looks warm and slightly orange against blue, and cool and slightly blue against orange. It is the same grey.
• A mid-grey looks light against black and dark against white. This is why you cannot judge a value against the white of your paper.
• A dull red placed next to a bright red looks brown. Place it next to grey and it looks vivid.

The practical consequences: never mix a colour in isolation and expect it to work in the picture, always test a mixed colour against the colours it will actually touch, and never judge a passage while the rest of the picture is still white paper.

Warm, cool, and the fact that every hue is both

Roughly, reds, oranges and yellows are warm; blues, greens and violets are cool. That is the version everyone knows and it is not very useful on its own, because every hue has warm and cool versions.

Hue	Warmer version	Cooler version
Red	Cadmium red, leaning orange	Alizarin crimson, leaning violet
Yellow	Cadmium yellow deep, leaning orange	Lemon yellow, leaning green
Blue	Ultramarine, leaning violet	Phthalo or cerulean, leaning green

This matters enormously for mixing, and Lesson 9 is built on it. A clean violet needs a red and a blue that both lean violet; mix a green-leaning blue with an orange-leaning red and you get mud, because you have quietly included all three primaries.

Temperature also does spatial work: warm colours tend to come forward and cool ones to recede, which is the colour half of aerial perspective from the last lesson. But value does far more of that work than temperature does, so do not rely on it to fix a flat picture.

Saturation is the dimension nobody controls

Students who have learned about contrast often make every colour as intense as it will go, and the result is exhausting and unfocused, because when everything shouts nothing is emphasised.

The professional habit is the reverse: keep most of the picture in a fairly narrow, fairly low saturation band, and spend your one or two genuinely saturated notes where you want the eye to land. Look at almost any painting you admire and measure this. The proportion of truly intense colour is usually small.

Van Gogh's Wheat Field with Cypresses, painted in 1889 and now at the Metropolitan Museum of Art, is often remembered as a riot of colour. Look at it carefully and it is organised: the field and sky occupy a limited range, and the picture works because of the relationship between a small number of decisions, not because every square centimetre is at maximum intensity.

Colour schemes, as scaffolding

Scheme	What it is	Effect
Monochrome	One hue, varied in value and saturation	Unified, calm; forces you to work with value
Analogous	Three or four neighbouring hues	Harmonious, easy; can become bland without a value plan
Complementary	Two opposite hues	Maximum tension; needs one of them to dominate
Split complementary	A hue plus the two neighbours of its opposite	Most of the punch, less crudeness
Triadic	Three hues equally spaced	Lively and hard to control; usually needs two of the three muted

Treat these as starting constraints, not as rules. Their real value is that choosing one in advance stops you reaching for a new tube every time you are stuck, which is the main cause of incoherent colour.

Optical mixing

Colours placed side by side in small enough marks blend in the eye rather than on the palette. Georges Seurat built an entire method on this, dotting unmixed colour so that the mixing happened in the viewer. The Metropolitan Museum holds his Circus Sideshow (Parade de cirque), of 1887 to 1888, a night scene built almost entirely from small separate touches.

The claim usually made for this, that optical mixture is more luminous than physical mixture, is only partly true, and Seurat's paintings are in fact often quite low in saturation up close. What is definitely true is that separated marks keep their identity: a green made from adjacent blue and yellow dots reads as a livelier green than the same green mixed flat, because it is not one thing. You can use this at any scale, with hatching, dry brush or coloured pencil, without adopting the whole system.

In short: two colours next to each other do something that the same two colours mixed cannot do.

Common misconceptions

• "You can mix any colour from red, yellow and blue." You cannot. The RYB set has a limited gamut; a clean violet or a clean turquoise is difficult or impossible from it, which is why printers use cyan and magenta.
• "Mixing complementaries gives a neutral grey." In theory. In practice it usually gives a brown or a slightly muddy neutral, and the exact result depends on the pigments. Mix them and see, then keep the swatch.
• "To darken a colour, add black." Black darkens and desaturates and often kills the colour. The complement, or a darker analogous hue, keeps it alive.
• "A colour has a fixed appearance." Chevreul's whole point is that it does not. Judge every colour in place, never in isolation.
• "More saturation means more impact." Impact comes from contrast, and contrast requires most of the picture to be less saturated than the part you want noticed.

The takeaway

Every colour has three dimensions, hue, value and saturation, and when colour is not working the fault is usually value. RYB is a school convention with a restricted gamut, CMY is what pigment mixing actually wants, and RGB is light, which is why screen colours often cannot be reproduced in paint. Complementaries maximise contrast when placed side by side and neutralise when mixed, and mixing towards the complement is how you grey a colour without deadening it. Chevreul's law of simultaneous contrast means no colour can be judged in isolation. Every hue has warm and cool versions, and mixing across that divide is the usual cause of mud. Control saturation deliberately: keep most of the picture restrained and spend the intensity where you want the eye. Choose a scheme in advance as scaffolding. And remember that two colours side by side do something two colours mixed cannot. Next lesson takes all of this to an actual palette.

Sources

1. Wikipedia contributors. (2024). Michel Eugène Chevreul. Wikipedia. en.wikipedia.org
2. Wikipedia contributors. (2024). Complementary colors. Wikipedia. en.wikipedia.org
3. Wikipedia contributors. (2024). RYB color model. Wikipedia. en.wikipedia.org
4. Seurat, G. (1887-88). Circus Sideshow (Parade de cirque) [Oil on canvas]. The Metropolitan Museum of Art, New York. metmuseum.org
5. Chevreul, M. E. (1839). De la loi du contraste simultané des couleurs. Paris: Pitois-Levrault.

For most of European painting history, the best blue came from lapis lazuli mined in Badakhshan, in what is now Afghanistan, ground and washed through a laborious process to yield ultramarine. It cost more than gold, which is why contracts sometimes specified how much of it a painter had to use and where. In 1824 a French industrial society offered a prize for a synthetic equivalent; two years later Jean-Baptiste Guimet claimed it. The blue that had been reserved for the Virgin's robe became a tube you can now buy for the price of a sandwich.

That history is worth knowing for a practical reason. Paint is not colour. Paint is a physical material with a price, a transparency, a drying behaviour and a lifespan, and mixing it well means knowing the material, not just the theory.

The case for a small palette

A rack of twenty-four tubes is worse than six. With twenty-four you hunt for the tube that matches; with six you have to mix, and mixing is what teaches you colour. It also produces harmony automatically, because every colour on the surface shares pigments with every other one.

The standard working set is the split-primary palette: a warm and a cool version of each primary, plus white. Seven items, and it will mix nearly everything you need.

Primary	Warm version (leans towards red or orange)	Cool version (leans towards blue or green)
Red	Cadmium red light	Alizarin crimson or quinacridone magenta
Yellow	Cadmium yellow deep, or yellow ochre as a cheap stand-in	Lemon yellow
Blue	Ultramarine, which leans violet	Phthalo blue or cerulean, which lean green
Plus	Titanium white	Optional: burnt sienna, which earns its place

Worth holding on to: to mix a clean secondary, use the two primaries that already lean towards it. Everything else in this lesson follows from that sentence.

Why that rule works

Say you want a clean violet. Violet sits between red and blue. So use the red that already leans blue, which is alizarin, and the blue that already leans red, which is ultramarine. The mixture contains almost nothing but red and blue, and the violet is clean.

Now do it wrongly. Take cadmium red, which leans orange, meaning it contains a suggestion of yellow, and mix it with phthalo blue, which leans green, meaning it also contains a suggestion of yellow. The mix now has red, blue and a double dose of yellow: all three primaries. All three primaries in a mixture is the definition of a neutral, and what comes out is a grey-brown.

Target	Clean mixture	Muddy mixture
Violet	Alizarin plus ultramarine	Cadmium red plus phthalo blue
Bright green	Lemon yellow plus phthalo blue	Cadmium yellow deep plus ultramarine, which gives olive
Vivid orange	Cadmium yellow plus cadmium red	Lemon yellow plus alizarin, which gives a dull salmon

Note that the muddy column is not useless. Olive green and dull salmon are extremely useful colours, and most of a picture is made of them. The point is to produce them on purpose.

Darks without black

Tube black is flat. It has no hue to relate to anything else, and used to darken a colour it drains the life out of it. Mix your darks instead.

• Ultramarine plus burnt sienna gives a deep near-black that you can tip warm by adding sienna or cool by adding blue. This one mixture will cover most of your dark needs and always harmonises, because both pigments are already in your other mixtures.
• A hue plus its complement, pushed a long way, gives a dark version of that colour rather than a grey version.
• A darker analogous hue: to darken a yellow, move it towards raw umber or a deep orange rather than towards black, which turns yellow into a sickly green.

Similarly, greys. A mixed grey made from your own primaries plus white will always sit correctly in the picture, because it is made of the same pigments as everything around it. A grey squeezed from a tube of neutral grey usually does not.

What white and black actually do

Adding white does two things at once, and only one of them is wanted. It raises the value, which is usually the intention, and it drops saturation, which usually is not. It also cools most colours: a red plus white does not become a pale red, it becomes pink, which is a different hue relationship.

So when you need something lighter and still intense, do not reach for white first. Try a lighter, more intense hue of the same family: to lighten a red, add a little orange or cadmium yellow before you add white. Reserve white for the genuinely pale passages.

In watercolour the question does not arise, because there is no white paint in the traditional method. The white is the paper, so you work light to dark and every layer must leave the light behind it. That single constraint is why watercolour is harder to correct than any other medium and why its practitioners plan so carefully.

Transparency, opacity and glazing

Behaviour	Typical pigments	What it is for
Opaque	Cadmiums, titanium white, cerulean	Covering, sitting on top, blocking in
Transparent	Alizarin, phthalo, quinacridone, most earth reds	Glazing: a thin layer that changes what is underneath

A glaze is a thin transparent layer over a dry one. It changes the colour beneath without hiding it, and because the light passes through and bounces back off the layer below, the result has a depth that flat opaque colour does not. The classic use is to warm or cool a whole area at once, or to unify a picture that has become patchy, by glazing a single thin colour over everything.

The rule when glazing in oil is fat over lean: each successive layer should contain slightly more oil than the one below it, or the upper layers will crack as they dry. In acrylic this does not apply, but a different limitation does: acrylic dries fast and dries slightly darker than it looks wet, so a colour mixed to match while wet will be wrong once dry. Mix a swatch, dry it with a hairdryer, and compare then.

How the medium changes the problem

Medium	Key behaviour	What to plan for
Watercolour	No white paint; light comes from the paper	Work light to dark; reserve your whites before you start
Gouache	Opaque watercolour; dries lighter than it looks wet	Mix slightly darker than the target; can be reworked when rewetted
Acrylic	Dries in minutes and slightly darker; waterproof once dry	Mix more than you need, work in layers, keep a wet palette
Oil	Dries in days; blends on the surface for hours	Soft transitions are easy; fat over lean; ventilation matters with solvents
Coloured pencil and pastel	Layered and optically mixed, not physically mixed	Build in thin layers; the paper tooth fills up and then accepts nothing more

Pigments do not last forever

Van Gogh painted The Bedroom in Arles in 1888 and made a second version in 1889 that is now at the Art Institute of Chicago. The walls in that painting read today as pale blue. Conservation research at the Art Institute has shown that they were originally violet: the red component was a light-sensitive lake pigment that has faded, leaving the blue behind. The painting we look at is a different colour from the one he made.

This is why paint tubes carry lightfastness ratings, usually a scale marked with roman numerals or asterisks, and why some brilliant cheap pigments are marked as fugitive. For studio exercises it does not matter. For work you want to keep, or submit, or sell, buy the higher rating and keep finished work out of direct sunlight.

Check your colour in black and white

The most useful habit in this lesson costs nothing. Photograph your painting and convert the image to greyscale. Every colour decision you have made now shows up as a value decision, and almost every failure of a colour picture turns out at that point to be a failure of value: two areas you thought were contrasting are in fact the same grey.

Bottom line: if the greyscale version reads clearly, the colour version will too. If it does not, no amount of hue adjustment will save it.

Common misconceptions

• "More tubes means more colours available." More tubes means less mixing and less harmony. Six pigments plus white will out-perform twenty-four for a beginner.
• "Mud is a mistake." Mud is a mixture containing all three primaries, and most of any picture is made of exactly that. The skill is producing it deliberately rather than by accident.
• "White lightens a colour." It lightens and desaturates and often cools it. To lighten while keeping intensity, move towards a lighter hue in the same family first.
• "Acrylic looks the same wet and dry." It dries noticeably darker. Always test a dried swatch before matching.
• "Any paint will last if you look after it." Lightfastness varies enormously by pigment. The walls of Van Gogh's Bedroom are the evidence.

Summing up

Paint is a material, not an abstraction, and mixing well means knowing what your pigments do. A split-primary palette of a warm and a cool version of each primary, plus white, mixes nearly everything, and the governing rule is to use the two primaries that already lean towards your target. Mixtures containing all three primaries turn neutral, which is where mud comes from and also where most of the useful colours in a picture come from. Mix your darks from ultramarine and burnt sienna rather than reaching for black, and mix your greys from your own pigments so they belong. White raises value but drops saturation and cools; watercolour removes the option entirely. Glazes change a colour without hiding it, and every medium has a behaviour you have to plan around, especially acrylic drying darker. Pigments fade, as the violet walls of Van Gogh's Bedroom demonstrate. And check every colour decision in greyscale, because most colour failures are value failures. Module 4 moves from what you put down to how you arrange it.

Sources

1. Wikipedia contributors. (2024). Ultramarine. Wikipedia. en.wikipedia.org
2. Wikipedia contributors. (2024). Primary color. Wikipedia. en.wikipedia.org
3. van Gogh, V. (1889). The Bedroom [Oil on canvas]. The Art Institute of Chicago. artic.edu
4. Getty. Publications and reports, including freely available conservation titles on pigments and painting materials. J. Paul Getty Trust. getty.edu

Set a jug, a lemon and a folded cloth on a table. Put the jug in the middle. Now draw all three carefully: correct proportions, clean ellipses, a full range of value from the white of the cloth to the dark under the jug. Two hours later you will have a drawing in which nothing is wrong and nothing happens.

This is the commonest serious failure in a student portfolio, and it is not a drawing failure. Everything Modules 2 and 3 taught can be present and the picture can still be dead, because accuracy decides what goes down and composition decides where. Let us take that still life apart and find the exact points where it died.

Diagnosing the dead still life

Four things went wrong, and they went wrong before the first mark.

1. The jug is centred. A subject in the exact middle of a rectangle divides the surrounding space into equal amounts on left and right. Equal amounts give the eye no reason to travel, so it lands on the jug and stops.
2. The three objects are roughly the same size. When nothing is clearly the largest, nothing is clearly the subject, and a viewer spends the first seconds hunting for one.
3. Everything is inside the frame with a margin. Three objects floating in the middle of a page with paper all round them read as specimens on a card, not as things in a place.
4. The gaps were never designed. The space between the lemon and the jug, and the space between the cloth and the bottom edge, were whatever happened. Both are shapes, and both are as visible as the objects.

Fixing any one of these takes about thirty seconds at the thumbnail stage and is nearly impossible two hours in. That asymmetry is the whole practical argument for composing first.

So what?: composition is not decoration applied to a finished drawing. It is the set of decisions that determine whether the drawing was worth making, and they are cheap to change only at the start.

The edge is part of the picture

The single fastest improvement available to most students is to stop treating the paper as a neutral background and start treating its four edges as active. Every shape you draw makes a second shape with the edge, and the eye reads both.

Three consequences follow. First, a form that touches or crosses an edge sits in a world that continues beyond the page, while a form with a margin all round it sits on the page. Second, a shape that comes very close to an edge without touching it creates a thin sliver of background that pulls attention like a splinter, so either commit to the crop or move away from it. Third, a horizontal placed at the exact half way point cuts the picture into two competing rectangles, which is why a horizon at the midline almost always feels weaker than one placed high or low.

You already know this from photographs of your own work. A photograph of a drawing with equal white margins looks like a document. Crop into it and it becomes an image.

Reading Bruegel's Harvesters

Pieter Bruegel the Elder painted The Harvesters in 1565, in oil on oak, at roughly 119 by 162 centimetres. It hangs in the Metropolitan Museum of Art in New York, and it belongs to a series on the times of the year of which five panels survive, three of them in Vienna. Look at how it is built.

A single pear tree stands slightly left of centre and runs the full height of the panel. It is the only large vertical in the picture and it does two jobs at once: it stops the eye from sliding straight off the left edge, and it divides the composition into a near half, where people are eating and sleeping, and a far half, where the wheat and the sea recede. The cut wheat runs as a pale diagonal wedge from the lower left towards the middle distance, and that wedge is the road your eye takes into the depth.

Now notice what Bruegel does not do. He does not put the tree in the middle. He does not put the horizon in the middle: it sits high, so that most of the panel is land and the picture is about work rather than sky. He does not make every figure the same size or the same distance apart. The sleeping man at the base of the tree is the largest single figure, and he lies at an angle across the direction of everything else, which is why you find him before you find anyone else.

Go to the Metropolitan's page for the painting, zoom in, and trace with a finger the path your eye takes. Most people go tree, sleeper, the eaters, then out along the pale wedge of cut wheat to the tiny figures at the far edge. That path was designed.

The rule of thirds, and what it is really for

Divide your rectangle into thirds horizontally and vertically. You now have four intersections. Put the thing that matters most on or near one of them, and put your main horizontal or vertical along one of the dividing lines rather than through the centre.

The phrase comes from the English painter and engraver John Thomas Smith, who set it out in Remarks on Rural Scenery in 1797 as a rule for balancing light and dark in a landscape. It has since been repeated so often that people either treat it as a law or dismiss it as a cliche. Neither is useful. It is a device that guarantees you will not centre things and will not split the picture in half, which are the two errors that killed the jug. Use it as a floor, not a ceiling: it prevents the worst outcome and it will never by itself produce a great composition.

Two cases where you should break it deliberately. Dead centre is powerful when you want confrontation and stillness, which is why formal portraits and religious icons use it. And an extreme edge placement, with the subject almost falling out of the frame, produces tension that thirds cannot give you. The rule exists so that you notice you are breaking it.

Hiroshige cuts the bridge in half

Utagawa Hiroshige's Sudden Shower over Shin-Ohashi Bridge and Atake, a woodblock print of 1857 from the series One Hundred Famous Views of Edo, is in the Metropolitan Museum's collection at about 34 by 24 centimetres. It does almost everything a European academic composition of the same date would forbid.

The bridge enters at the left edge and leaves at the right edge. It is not contained. It is sliced by the frame at both ends, so the viewer is placed not in front of the scene but inside it, standing on the riverbank in the rain. The figures on the bridge are cropped by its parapet. The rain is a set of straight parallel lines ruled across the whole print, including across the figures and the far shore, which no observed shower actually looks like and which reads instantly as rain. The far bank is a flat dark band with no detail at all. Vincent van Gogh copied the print in oil in 1887, and the copy is in the Van Gogh Museum in Amsterdam.

Set that against the Bruegel. Bruegel builds a deep, populated world you travel into. Hiroshige flattens everything into a few strong shapes and crops hard. Both are excellent compositions, and they share only one thing: nothing important is in the middle and nothing was left to chance.

The principles, as a checklist you can run

The elements of art are what you put down: line, shape, form, value, colour, texture, space. The principles of design are what you do with them. Here they are as questions you can ask of a thumbnail in under a minute.

Principle	The question to ask	The failure it catches
Dominance	What is the largest, darkest or most isolated thing? Is that what I want the subject to be?	Three objects of equal weight and no subject
Balance	If the picture were a tray, would it tip? Is the balance symmetrical, or a small dark thing counterweighting a large pale one?	Everything heavy on one side by accident
Movement	Where does the eye enter, where does it go next, and what stops it leaving?	A path that runs straight off the edge and does not come back
Rhythm and repetition	Is any shape, angle or interval repeated? Is it repeated with variation, or identically?	Evenly spaced identical marks, which read as wallpaper
Contrast	Where are the lightest and darkest points, and are they near what matters?	Maximum contrast in a corner, dragging the eye away from the subject
Unity and variety	Does everything belong to one picture, and is there enough difference to hold attention?	Either a jumble, or a beautifully unified nothing
Scale and proportion	How big is the subject relative to the format, and why that size?	A small subject stranded in a large empty rectangle by default

Do not run all seven every time. Run dominance, movement and contrast on every thumbnail, and bring in the others when something feels wrong and you cannot name it.

Notan: testing a composition in two values

Notan is a Japanese term for the interplay of light and dark, and it entered Western art teaching through Arthur Wesley Dow's book Composition in 1899, which shaped a generation of American teaching, including Georgia O'Keeffe's. The exercise is simple and unusually effective.

Take a small rectangle, about seven centimetres by ten, and fill every part of it with either black or white. No greys, no lines, no detail. You must decide, for every square millimetre, whether it belongs to the light family or the dark family. Then look at the result from two metres away with your eyes half closed.

What this exposes is whether the composition works as an arrangement of shapes, independent of subject, drawing skill and colour. Most weak compositions collapse immediately: the dark shapes turn out to be scattered evenly, or the subject dissolves into a background of the same value, or the whole rectangle is fifty percent grey noise. Strong compositions survive, because a strong composition is a good arrangement of a small number of shapes.

Do three or four notan studies of the same subject with different crops and different decisions about what belongs to the dark. It takes ten minutes for all four and it will change which drawing you spend the next two hours on.

The upshot: if a composition does not work in two values at postage stamp size, no amount of rendering will save it. If it does work, the rendering has somewhere to go.

Format is a decision

Most students draw on whatever shape the pad happens to be. The proportion of the rectangle is itself a compositional choice, and choosing it deliberately costs nothing.

A tall narrow format forces verticality and compresses space sideways, which suits a standing figure or a single tree. A wide format pulls the eye horizontally and suits landscape and narrative, which is why Bearden's The Block, which you will meet in Module 5, is eighteen feet long. A square resists both and produces stillness, because neither dimension wins. And a shaped or cropped format, such as a long strip cut from a larger sheet, can rescue a drawing that is failing: crop it and see.

Before you start, cut two L shaped pieces of card. Overlap them to make an adjustable rectangle and hold it up to your subject. Slide them until the arrangement inside the opening looks right, then measure the ratio you have arrived at and draw on that ratio. This is the same viewfinder tool from the sighting lesson, used for a different purpose.

Common misconceptions

• "Composition is what you fix at the end." It is what you decide at the start. Two hours of rendering cannot move a centred jug.
• "The rule of thirds is the secret." It is a floor that prevents the two worst placements. It has never on its own made a picture good.
• "Empty areas are wasted space." The empty areas are shapes, they have weight, and a large quiet area is often what gives the busy area its force.
• "Balance means symmetry." Symmetry is one kind of balance and the most static. A small dark shape can balance a large pale one, which is more interesting and harder.
• "If I crop the subject the viewer will not know what it is." They will, and cropping is what puts them inside the scene rather than in front of it. Hiroshige cut the bridge at both ends.

What to remember

An accurate drawing can be a dead picture, and the death happens in the first thirty seconds, when the subject is centred, the objects are given equal weight, everything is floated inside a margin and the gaps are left to chance. The edges of the paper are part of the picture: crossing them puts the viewer inside the scene, and a shape that almost touches an edge creates a sliver that steals attention. Bruegel's Harvesters at the Metropolitan builds a route into deep space with one off centre tree and a diagonal wedge of cut wheat; Hiroshige's rainstorm print, in the same collection, slices its bridge at both edges and flattens the far bank into a single band. Both are composed to the millimetre. Use the rule of thirds as a floor, the seven principles as questions rather than laws, and a two value notan study to test whether the arrangement stands up before you spend two hours on it. Choose the shape of the rectangle deliberately. Next lesson turns to the marks themselves, and to why the same drawing in charcoal and in pen is not the same drawing.

Sources

1. Bruegel, P., the Elder. (1565). The Harvesters [Oil on wood]. The Metropolitan Museum of Art, New York. metmuseum.org
2. Hiroshige, U. (1857). Sudden Shower over Shin-Ohashi Bridge and Atake, from One Hundred Famous Views of Edo [Woodblock print]. The Metropolitan Museum of Art, New York. metmuseum.org
3. Tate. Composition [Art term]. Tate, London. tate.org.uk
4. Wikipedia contributors. (2024). Rule of thirds. Wikipedia. en.wikipedia.org
5. Wikipedia contributors. (2024). Notan. Wikipedia. en.wikipedia.org

In July 1888, near Arles, Vincent van Gogh cut reeds from the bank of a ditch, sharpened them into pens and drew a wheat field with them. The sheet is small, about 24 by 32 centimetres, and it is in the Metropolitan Museum of Art, catalogued as reed pen and logwood ink over graphite on wove paper. He could have bought a steel nib. He deliberately used a tool that blunts as you work, holds very little ink, and produces a dry scratchy stroke at the end of every dip.

That is the whole subject of this lesson. Van Gogh did not choose the reed because it was authentic or cheap. He chose it because a reed makes a particular mark, wide then thin then broken, and a field of ripening wheat in a hot wind is made of exactly that kind of mark. The material was an argument about the subject.

What a mark records

Every mark on a page carries four pieces of information whether you intend it or not: how hard you pressed, how fast you moved, what the tool was, and what the surface did to it. A viewer reads all four without knowing they are doing so, which is why a drawing made with the wrong material feels wrong before anyone can say why.

Two students draw the same crumpled paper bag. One uses a 2H pencil on smooth cartridge paper, keeps the pressure even, and produces a grey, sealed, metallic object. The other uses compressed charcoal on rough paper, varies pressure hard, and produces something that reads as dry, brittle and creased. Neither is more accurate. They are answers to different questions, and only one of the students chose.

Worth holding on to: the material is not a delivery system for an image you already have. It is part of what the image says.

The dry media, compared

Material	Darkest value	Erases?	What it is good at	What it fights you on
Graphite pencil, H grades	Light grey	Fully	Fine detail, construction lines, controlled edges	Grooves the paper if pressed; will not go dark
Graphite pencil, B grades	Mid to dark grey	Mostly	Value range with control; the everyday workhorse	Shines under light, so blacks read grey
Graphite stick or powder	Dark grey	Partly	Covering large areas fast; broad soft passages	Very hard to keep clean; smudges everywhere
Vine or willow charcoal	Mid grey	Almost completely, with a cloth	Fast searching, big changes, blocking in	Falls off the paper; cannot hold a dark
Compressed charcoal	True black	Reluctantly	Deep darks, decisive commitment, scale	Nearly permanent once burnished in
Conté crayon	Dark, slightly warm	Barely	Granular tone on textured paper; earth colours	Little fine detail; the tooth controls you
Coloured pencil	Dark, medium saturation	Partly	Slow layered colour, transparency, precision	Wax bloom and slow going over large areas

Read the last column first. The reason a drawing goes wrong is usually written there.

Graphite grades are about clay, not just darkness

H stands for hard, B for black, F for fine, and the scale runs from about 9H to 9B with HB and F in the middle. What changes along that scale is the proportion of clay to graphite in the core: more clay means a harder, lighter, more durable point, and less clay means a softer, darker, faster wearing one.

Two practical consequences follow, and they explain most pencil frustration. First, a hard pencil does not just draw lighter, it draws thinner and holds a point through a whole passage, which is why construction lines and hatching in a small drawing want an H or an HB. Second, a hard pencil pressed hard does not go darker. It embosses. It cuts a groove into the paper that no eraser will lift and that will show as a shiny white line through any tone you lay over it later. If you catch yourself pressing to get darker, you have the wrong grade in your hand.

The other limit is the shine. Graphite is a mineral that reflects light, so a heavily worked graphite black looks grey and slightly silver when the drawing is tilted towards a window. This is why photographs of graphite drawings so often disappoint, and why a drawing that needs a genuine black usually needs charcoal, ink or a dark Conte instead.

Charcoal, and the two things that are rarely said

Charcoal is carbonised wood: willow and vine charcoal are simply charred sticks, while compressed charcoal is charcoal powder held together with a gum binder and pressed into a stick.

The first thing rarely said is that vine and compressed charcoal are not grades of the same material, they are different materials with different jobs. Vine charcoal sits on the surface, comes off almost completely with a chamois or a swipe of the hand, and is therefore the correct tool for the first forty minutes of a large drawing, when you are still moving things. Compressed charcoal grips, goes genuinely black, and is very hard to remove; it is the tool for the last twenty minutes. Students who fight their drawings usually started with the wrong one.

The second is that the eraser is a drawing tool. A kneaded eraser can be shaped to a point, pressed to lift a highlight without disturbing anything around it, or dragged to pull a light line out of a dark field. In charcoal you do not draw darks onto white paper; you tone the whole sheet mid grey, then take light out and push dark in. That reversal is the single largest jump in a beginner's charcoal work.

Seurat, Conté crayon, and the tooth of the paper

The Metropolitan Museum holds Georges Seurat's Aman-Jean, a portrait of the painter Edmond François Aman-Jean made in 1882 to 1883, about 62 by 47 centimetres, in Conté crayon on Michallet paper. Look at a detail of it and there is no line anywhere. There is no hatching either. There is a continuous granular fog out of which a head emerges.

That effect is not a technique so much as a collaboration with a surface. Michallet is a laid paper: during manufacture it is formed on a screen of parallel wires, which leaves a fine ribbed texture of raised and sunken lines. A crayon dragged across it deposits pigment on the raised ridges and skips the pits, so every stroke is automatically broken into a field of dark specks and white gaps. Press harder and the pits begin to fill; press lightly and the white dominates. Seurat controlled value almost entirely by how much of the paper he allowed to stay white.

Try the same crayon on smooth hot pressed paper and the magic is gone, because there is nothing for it to skip over. The lesson generalises: a textured or laid surface converts pressure into value, while a smooth surface converts pressure into darkness. Choose accordingly.

The wet media, compared

Material	Behaviour	Lightfastness	Best use
India ink, carbon based	Dense black, waterproof when shellac bound, permanent on contact	Excellent; carbon does not fade	Anything you need to survive, and any work to be scanned
Dye based writing inks	Fluid, brilliant colour, dilutes beautifully	Poor; many fade badly in light	Sketchbook work and studies, not finished pieces for display
Dip pen, steel nib	Line varies with pressure; flexible nibs swell enormously	Depends on ink	Precise varied line, cross hatching
Reed or bamboo pen	Blunt, thirsty, produces dry broken strokes as it empties	Depends on ink	Energetic textural drawing at speed, as in the van Gogh
Brush and ink	Line from a hair's width to a broad sweep in one stroke	Depends on ink	Gesture, calligraphic figure work, large simple shapes
Ink wash	Diluted ink laid in flat or graded areas	Depends on ink	Value structure under a line drawing

Van Gogh's sheet is catalogued as logwood ink, a dye extracted from a Central American tree. Dye based inks are among the least lightfast materials in a studio, which is exactly the lightfastness problem you met with pigments in Module 3. If a drawing matters, use carbon.

Paper is a material, not a background

Four properties of paper change your drawing more than most people expect.

• Weight, given in grams per square metre. Cartridge paper for pencil is around 130 to 150 gsm. Anything wet needs 300 gsm or it will cockle into waves that never flatten.
• Surface. Hot pressed is smooth and suits pen and fine pencil. Cold pressed, also called NOT, has a moderate tooth and is the general purpose choice. Rough has a pronounced texture that breaks up every stroke, which is wonderful for charcoal and unusable for detail.
• Tooth, the microscopic roughness that grabs dry media. No tooth means charcoal and pastel simply will not stick, which is why they slide off glossy paper.
• Sizing, a gelatine or synthetic treatment that stops liquid soaking in. Unsized paper drinks ink and blots; heavily sized paper lets washes sit and move on the surface.

Then there is survival. Cheap paper is made from wood pulp containing lignin, which oxidises and turns the sheet yellow and brittle within a few years. Acid free or pH neutral paper does not, and the difference costs very little. Work you intend to photograph for a portfolio next spring can be on anything; work you want to still exist at twenty five should be on acid free stock.

In short: the same gesture with the same tool on three different papers gives three different drawings, and the paper is chosen before the tool.

Building a mark vocabulary you can actually use

Most students own about four marks: an outline, a hatch, a scribble and a smudge. A working artist has perhaps twenty and can reach for one on purpose. The way to acquire them is not to read about them but to make a labelled sheet and keep it.

Marks worth having: hatching at a consistent angle; cross hatching at two or three angles; contour hatching that follows the curve of a form; stippling; scumbling in small circles; broad side of the stick; blunt edge; sharp edge; dry brush; wet into wet; a lifted mark taken out with a kneaded eraser; a burnished mark pressed flat with a paper stump; a broken mark made by drawing over a textured surface underneath the paper; and an incised line drawn hard with a dead ballpoint before you lay tone over it, so it stays white.

Frottage, the last one but with texture, is worth trying today: put a leaf, a coin or a piece of wood grain under your sheet and rub the side of a soft pencil over it. Max Ernst built a whole body of work on that in the 1920s. It costs nothing and gives you a mark you did not invent.

Fixing, storing and the slow ruin of work

Charcoal, soft pastel and graphite all sit loose on the surface and will transfer to whatever they touch. Workable fixative locks a layer while allowing you to draw over it; final fixative is heavier and is for the end. Two warnings. Fixative darkens charcoal slightly and can flatten the lightest passages, so test on a scrap of the same paper before spraying a finished drawing. And spray outdoors or by an open window, in thin passes from about thirty centimetres, never a single wet coat, which will cause the surface to bloom.

Store drawings flat, interleaved with sheets of clean paper, out of sunlight. A portfolio of rolled drawings is a portfolio of drawings that will not lie flat when it matters. The lesson on documentation later in this course assumes your work still exists and is not curled, and that assumption is worth protecting now.

Common misconceptions

• "A softer pencil is just a darker pencil." It is a different clay ratio: softer means darker, blunter, faster wearing and less able to hold an edge. Pressing an H grade harder embosses the paper instead.
• "Charcoal is charcoal." Vine charcoal removes almost completely and cannot make a black; compressed charcoal makes a black and barely removes at all. They are for different halves of the same drawing.
• "Erasing is fixing a mistake." In charcoal the eraser is a drawing tool. You tone the sheet, then take the light out.
• "The paper is just what you draw on." Seurat's granular tone is made by the ribs of a laid paper, not by his hand. The same crayon on a smooth sheet gives a different drawing.
• "Ink is permanent." Carbon based India ink is. Many coloured and dye based inks fade noticeably in a few years of daylight.

Looking back

Van Gogh cut a reed from a ditch because a reed makes a dry broken stroke and a wheat field in a hot wind is made of dry broken strokes. Every mark you make records pressure, speed, tool and surface, and a viewer reads all four. Graphite grades differ by clay content, so a hard pencil pressed hard embosses instead of darkening, and graphite's shine means a true black usually needs charcoal or ink. Vine charcoal is for the searching stage and compressed for the committing stage, and in both the eraser is a drawing tool rather than a correction. Seurat's Aman-Jean at the Metropolitan shows what a surface contributes: the ribs of Michallet paper break every stroke into specks, so pressure becomes value. Choose paper for weight, surface, tooth, sizing and, if the work should survive, freedom from acid. Build a labelled sheet of twenty marks so you can reach for one on purpose. Next lesson takes the mark somewhere else entirely: onto a block or a plate, where you cut it once and print it many times.

Sources

1. van Gogh, V. (1888). Wheat Field [Reed pen and logwood ink over graphite on wove paper]. The Metropolitan Museum of Art, New York. metmuseum.org
2. Seurat, G. (1882-83). Aman-Jean (Portrait of Edmond François Aman-Jean) [Conté crayon on Michallet paper]. The Metropolitan Museum of Art, New York. metmuseum.org
3. Wikipedia contributors. (2024). Charcoal (art). Wikipedia. en.wikipedia.org
4. Wikipedia contributors. (2024). India ink. Wikipedia. en.wikipedia.org
5. Wikipedia contributors. (2024). Acid-free paper. Wikipedia. en.wikipedia.org

Rembrandt's Hundred Guilder Print, made around 1648 and held by the Metropolitan Museum of Art, is a plate of about 28 by 39 centimetres worked with etching, engraving and drypoint together. The name is a nickname: it records the sum an impression is said to have fetched in the seventeenth century, a startling price for a piece of paper. That is the point of the story. The image was worth that much because it existed in more than one copy and could travel, and because Rembrandt could go on changing the plate between printings.

Printmaking is the branch of art in which you do not make the picture. You make the thing that makes the picture, and then you make the picture from it, more than once. That one step of indirection changes everything about how you plan.

Four families, sorted by where the ink sits

Every printmaking process in the world belongs to one of four families, and the classification is not historical or stylistic. It is purely about which part of the surface holds the ink.

Family	Where the ink sits	What you cut or draw	Processes
Relief	On the raised surface	You cut away everything that should stay white	Woodcut, linocut, wood engraving
Intaglio	In the incisions, below the surface	You cut or bite the lines that will print black, then wipe the surface clean	Engraving, etching, drypoint, aquatint
Planographic	On a flat surface, chemically	You draw with grease on stone or plate; grease takes ink, damp stone repels it	Lithography
Stencil	Passes through openings	You block the areas that stay blank and push ink through the rest	Screenprint, stencil

Relief and intaglio are exact opposites, and confusing them is the commonest early error. In a linocut, the gouge makes white. In an etching, the needle makes black. If you sit down to cut a linocut thinking of the tool as a pencil, you will produce a photographic negative of what you intended.

The core of it: before you make a single mark on a block or plate, say out loud whether your tool is making the blacks or making the whites. Everything else follows from that.

A linocut from blank block to signed print

Lino is the accessible one: cheap, needs no press, and forgiving enough for a kitchen table. Here is the whole procedure.

1. Design in two values first. Make a notan study, exactly as in the composition lesson. A linocut has no greys, only ink and paper, so if the design does not work in black and white it will not work at all. Anything that looks like grey in a print is an illusion made by hatching or texture.
2. Transfer, reversed. Draw or trace your design onto the block, remembering that the print will be a mirror image. Any text must be cut backwards.
3. Ink the block black before cutting. Roll or brush a thin coat of black ink or paint over the lino and let it dry, then draw on it in white pencil. Now every cut reveals pale lino, so the block looks like the print as you work. This single habit removes most beginner confusion.
4. Cut, working from small tools to large. Start with the fine V gouge for outlines and detail, then clear open areas with the wide U gouge. You can always cut more away; you cannot put lino back.
5. Cut away from your hand, every time. Use a bench hook, a simple board with a lip at each end that braces the block against the table edge, so the block cannot slide. Lino tools slip, and they slip in the direction you are pushing. Never hold the block in the hand you are cutting towards.
6. Ink with a roller. Roll a small amount of block printing ink out on a smooth surface until it makes a fine even hiss, then roll it onto the block in several directions. Too much ink fills the fine cuts and turns detail into a blur.
7. Print by hand. Lay paper on the block, hold it still, and burnish the back firmly with the bowl of a wooden spoon or a smooth pebble, working systematically so nothing is missed. Peel a corner back to check before committing.
8. Pull a proof, then correct. The first pull is a proof, not a print. Look at it, mark on it what needs cutting away, cut, and pull again. Two or three rounds of this is normal and is where the print actually gets made.

The reversal that catches everybody

Anything printed by pressing paper onto an inked surface comes out mirrored left to right. Every printmaker has cut a beautiful block with a word on it and discovered the word backwards on the paper.

Two habits prevent it. Hold your drawing up to a mirror, or photograph it and flip the image on your phone, and work from the flipped version. And whenever direction carries meaning, such as a figure looking towards something, decide at the design stage which way the printed figure should face and reverse before you transfer, not after you cut.

Registration: getting the second colour in the right place

The moment you print more than one colour, you have a registration problem: the second layer has to land in a known relationship to the first. The simplest solution needs no equipment.

Tape a sheet of card to the table and draw two rectangles on it: one the size of the block, one the size of the paper, sharing a corner. Now every time you print, the block goes in one rectangle and the paper in the other, and every impression lands in the same place. This is a registration jig, and it costs a piece of cardboard and four minutes.

Then leave yourself room to be slightly wrong. A hairline gap between two colours reads as a mistake; a deliberate overlap of a millimetre or two, where the darker colour prints over the lighter, reads as intended. Design for the overlap and small drift will not show.

The reduction print, where the block destroys itself

Now change one input and watch the whole procedure invert. In a reduction print, you use a single block for every colour, cutting more away between each printing. Picasso made colour linocuts this way from the late 1950s, and the technique is sometimes called the suicide print, because by the end the block cannot make the image again.

The sequence, for a three colour print:

1. Cut away only what must stay the white of the paper. Print the lightest colour, say yellow, on your whole edition at once.
2. Let it dry. Now cut away everything that should stay yellow. Print the middle colour, say red, over every sheet. Where red covers yellow you get orange, so plan for the mixture rather than the tube colour.
3. Cut away everything that should stay red or orange, and print the darkest colour last.

Three consequences follow, and they change how you work. First, you must print the entire edition at every stage, because you cannot go back: once the yellow stage is cut away, no further yellow impressions exist. Second, print more sheets than you need at stage one, because misprints at stage three are unrecoverable. Third, the colours mix by overprinting, so a reduction print is a colour mixing exercise in the sense of Module 3, run in transparent layers rather than on a palette.

This is the clearest example in the whole course of a process where the planning is the work. A reduction print made without a written plan of what is cut at each stage will fail, not because of skill, but because the information needed at stage three no longer exists at stage three.

Monotype: the print that refuses to be an edition

At the other extreme sits the monotype. You paint or roll ink onto a smooth unmarked plate of glass, plastic or metal, work into it with brushes, rags and cotton buds, and take a single impression. Because nothing is cut or etched, the plate holds no permanent image, and the second pull, called a ghost, is faint and different.

Monotype is worth a fortnight of anyone's investigation because it is fast, it is impossible to be precious about, and it produces marks you cannot make any other way: wiped, blotted, dragged, lifted with a fingertip. A trace monotype is even simpler. Roll ink onto a plate, lay paper on top face up, and draw on the back of the paper with a pencil. Wherever you press, ink transfers, and the resulting line has a soft velvety quality with faint marks from where your hand rested.

Numbering an edition, and doing it honestly

The conventions are simple, and following them signals that you know what you are doing.

• Sign and number in pencil, below the image, on the paper margin. Number at the left, title in the middle, signature at the right.
• A number like 12/30 means the twelfth impression of an edition of thirty. The number is not a quality rating.
• An edition is a fixed number of impressions of the same image, printed and then closed. Marking prints 1/30 and then printing fifty is dishonest, and it is the reason collectors care about the convention at all.
• AP means artist's proof, an impression kept outside the numbered edition, traditionally a small proportion of it.
• A trial proof or state proof is a pull taken while the block is still being changed, and it is often the most interesting sheet you own. Keep them. They are exactly the process evidence your portfolio needs.

Remember: in printmaking the failures are documented automatically. Every state of the block leaves a proof, and those proofs make the thinking visible in a way a finished drawing never can.

What repetition does to an image

Two works, both about the fact of multiplicity, pulling in opposite directions.

Käthe Kollwitz lost her younger son Peter in the first months of the First World War. Her portfolio Krieg, or War, appeared in 1923 as a set of woodcuts, and the Metropolitan Museum holds an impression of Widow II from it. The image is cut hard, almost entirely in black masses, with the wood's coarse grain left visible. She chose the medium because prints are cheap, travel, and reach people who do not go to galleries. Multiplicity there is a political tool.

Andy Warhol's Marilyn Diptych, made in 1962 and held by Tate in London, is acrylic and silkscreen ink on two canvases. Marilyn Monroe had died that August. On the left panel her face repeats in flat colour; on the right it repeats in black and white, and across that panel the printing degrades until the last impressions are smeared and half gone. The subject of the work is the repetition itself: an image reproduced until it wears out. That is a comment only a printmaking process could make.

When you make a print, ask which of these you are doing. Are you multiplying because you want the image in many hands, or because the repetition is part of the meaning, or simply because you like the mark that cut lino gives? All three are legitimate. Not knowing which is not.

Common misconceptions

• "A print is a copy of an artwork." A reproduction is a copy. An original print is made from a matrix created for that purpose, and every impression is the artwork rather than a picture of it.
• "In a linocut the gouge draws the lines." The gouge makes the whites. Cut a line and it prints as a white line through black.
• "Lower numbers in an edition are better." 2/30 is not superior to 27/30. The number records position, not quality.
• "I can go back and print more of the yellow stage later." Not in a reduction print. The yellow stage was cut away, permanently, which is why the whole edition is printed at every stage.
• "More ink gives a richer print." Too much ink floods the fine cuts and destroys the detail you spent an hour making. Roll it thin enough to hiss.

The short version

Printmaking puts one step between you and the image: you make the matrix, and the matrix makes the print. The four families are separated by where the ink sits, on the surface for relief, in the incisions for intaglio, chemically on a flat plane for lithography, and through openings for screenprint, and relief and intaglio are exact opposites, so your first decision is whether your tool is cutting whites or drawing blacks. A linocut runs from a two value design, through a reversed transfer onto a block inked black so the cuts read as light, to careful cutting away from your hand with a bench hook, thin ink, hand burnishing, and proofs that you correct and pull again. Everything prints mirrored. A cardboard jig solves registration and a deliberate overlap hides drift. The reduction print inverts the logic entirely by destroying the block as it goes, which is why the entire edition is printed at each stage and why the plan is the work. Monotype gives up the edition for marks nothing else makes. Number honestly, keep your trial proofs as process evidence, and know whether you are multiplying for reach, as Kollwitz did, or because the repetition is itself the subject, as in Warhol's Marilyn Diptych at Tate. Module 5 opens the image out further, starting with the scissors.

Sources

1. Rembrandt van Rijn. (1648). The Hundred Guilder Print [Etching, engraving and drypoint]. The Metropolitan Museum of Art, New York. metmuseum.org
2. Kollwitz, K. (1922-23). Widow II, from the portfolio War (Krieg) [Woodcut]. The Metropolitan Museum of Art, New York. metmuseum.org
3. Warhol, A. (1962). Marilyn Diptych [Acrylic paint and silkscreen ink on canvas]. Tate, London. tate.org.uk
4. Tate. Linocut [Art term]. Tate, London. tate.org.uk
5. Tate. Edition [Art term]. Tate, London. tate.org.uk

Romare Bearden's The Block is four feet high and eighteen feet long: six panels of cut and torn paper, photographs, foil, watercolour and ink, laid on Masonite. He made it in 1971 and it is at the Metropolitan Museum of Art. Stand at one end and it is a Harlem street in elevation, brownstones and a church and a barber shop and a funeral parlour, all frontal, all pressed flat like a stage set. Walk along it and the flatness gives way. Windows open into interiors. A cut out photograph of a real face looks out of a window built from a magazine advertisement. A whole life is going on inside a building drawn as a rectangle.

Bearden could draw. He chose scissors instead, and the reason is the subject of this lesson.

What a cut edge does that a drawn edge cannot

A drawn edge is continuous with everything around it. Same hand, same pencil, same paper. However hard you push, a drawing is one substance describing many things.

A cut edge is a join between two substances that were never together. A photograph of a face next to a piece of wallpaper next to a torn strip of newspaper: three different resolutions, three different eras of printing, three different origins, held on one surface. The viewer sees the join, and the join carries meaning that neither piece had alone.

This is why Bearden's faces are photographic while his buildings are flat coloured shapes. The mismatch is the content. Real people, in a stylised street, in a city that flattens them into facades. Nothing about that could be said in a single continuous medium without saying it much more slowly.

What matters here: collage is not a texture technique. It is a way of putting two unlike things next to each other and making the gap between them do the work.

Where it began, and what was at stake

In May 1912 Pablo Picasso glued a piece of printed oilcloth onto an oval canvas, framed the whole thing with a length of rope, and called it Still Life with Chair Caning. The oilcloth was machine printed with a pattern imitating woven cane. It is in the Musée Picasso in Paris and it is usually described as the first modern collage.

Notice what is actually radical there, because it is not the glue. Every other painter of 1912 who wanted chair caning in a picture would have painted it. Picasso used a printed imitation of caning, which is a picture of caning made by a machine, and put it in a painting. He introduced a second kind of image, made by someone else for another purpose, into a work of art. Everything in this lesson follows from that.

Kurt Schwitters took it further from 1919, building works from tram tickets, sweet wrappers, bus timetables and wire, material he picked up in the street in Hanover. He called the whole enterprise Merz, a fragment cut out of the word Kommerzbank. Tate holds a number of his works. The claim underneath them is that discarded printed matter is a legitimate material for art, which was not obvious then and made a great deal possible afterwards.

The political cut: photomontage

Photomontage is collage made specifically from photographs, and in Berlin around 1919 it became a weapon. Hannah Höch's Cut with the Kitchen Knife Dada Through the Last Weimar Beer-Belly Cultural Epoch of Germany, made in 1919 to 1920, is in the Nationalgalerie in Berlin. It is packed with heads of politicians, generals, dancers and machinery, cut from the illustrated press and reassembled so that a general's head sits on a dancer's body.

Two things make it work, and both are technical decisions you can copy. First, scale is deliberately wrong: a tiny head next to an enormous wheel, so no coherent space can form and the eye cannot settle. Second, the sources are recognisable. These were faces that readers of German newspapers knew. Cutting a known face out of its context and gluing it into another is an act performed on a real image, and the viewer feels it.

If you want your own photomontage to have force rather than merely look busy, decide first what you are cutting out of what. A collage assembled from pleasant textures found in a magazine is decoration. A collage assembled from a specific set of images, chosen because of where they came from, is an argument.

Matisse cuts directly into colour

By 1953 Henri Matisse was in his eighties and largely confined to bed. He had assistants paint sheets of paper with gouache in flat single colours, and he cut shapes out of them with scissors, directing where the pieces should be pinned on the wall until the arrangement was right. The Snail, made that year and held by Tate in London, is one result: a very large work, close to three metres in each direction, made of torn and cut coloured rectangles spiralling loosely around a centre.

His own phrase for the method was drawing with scissors, and it names something exact. Cutting a shape out of a coloured sheet gives you the contour and the colour in one action, with no line separating them. There is no outline anywhere in The Snail. Each shape is bounded only by where the colour stops.

The working method matters as much as the result. Nothing was fixed until it was right. Shapes were pinned, moved, recut, moved again over days. That is a version of the iteration you have been practising since Lesson 1, and it is the strongest practical argument for collage in a school portfolio: unlike a drawing, a collage can be entirely rearranged five minutes before you commit, so the composition can be found rather than guessed.

Cut edges against torn edges

Edge	How it reads	Use it when
Scissor cut	Deliberate, clean, slightly hard; the shape looks chosen	You want a shape to be read as a shape, as in Matisse
Knife cut against a straight edge	Mechanical, architectural, impersonal	Buildings, grids, anything that should feel manufactured
Torn with the printed side up	A soft feathered edge showing the white core of the paper	You want the join to soften, or the material to look aged and handled
Torn against a ruler	Straight but ragged; controlled damage	Long horizontals that should not look printed
Burned, sanded or abraded	Damaged, historical, unstable	Rarely, and never merely for effect

Use two kinds of edge in one piece and the difference between them becomes readable. Use five and it becomes noise.

Adhesives, supports, and what fails within a year

Collage has a reputation for falling apart, and it is deserved when the materials are chosen carelessly.

• Glue stick is convenient and temporary. Edges lift within months. Acceptable for sketchbook work and for trying arrangements, not for anything you want to photograph next spring.
• PVA, white wood or craft glue, thinned slightly with water and brushed thinly over the whole back of a piece rather than dotted at the corners, is the practical default. Buy an acid free version if you can. Dotting glue at the corners is the single commonest cause of curling.
• Wheat starch paste is what conservators use, because it is reversible with water and does not yellow. It takes ten minutes to cook and is worth knowing about.
• Acrylic medium glues and seals at once, and can be brushed over the top as well as underneath, which locks everything down. It changes the surface sheen, so test it.
• Rubber cement and cellophane tape both stain the paper brown and both let go. Avoid.

Then the support. Paper is not strong enough for a heavy collage: as the glue dries it shrinks and the sheet buckles into a shape no press will flatten. Work on card, mount board or a rigid panel, which is what Bearden did with Masonite. If you must use paper, tape all four edges down to a board with gummed paper tape while you work and cut it free at the end, exactly as watercolourists stretch paper.

Bottom line: choose the support before the imagery. A brilliant collage on 120 gsm cartridge paper is a brilliant collage that is buckling by the time you photograph it.

Mixed media: combinations that behave

Mixed media simply means more than one material in one work, but the combinations are not all equally well behaved.

Combination	What happens	Watch for
Wax crayon or oil pastel under watercolour	Wax resists water; the marks stay clean under a wash	Nothing will stick over the waxed area later
Ink over dry acrylic	Sits crisply on the sealed surface	Non waterproof ink will lift if you wash over it
Charcoal over acrylic or gesso	Gesso gives excellent tooth and takes charcoal beautifully	It must be fixed, or it will move for years
Collage under paint	Paint unifies disparate scraps into one surface	Unifying too much removes the reason you used collage
Acrylic over oil	Fails. Acrylic will not bond to an oily surface and eventually flakes	Always the other way round: oil over acrylic is fine
Photocopy or inkjet print in collage	Cheap, fast, and lets you change the scale of anything	Dye based inkjet ink runs when glue is brushed over it; test first

The photocopier deserves a line of its own. Enlarging a small image until the halftone dots become visible, or copying a copy of a copy until it degrades, is a genuine mark making method and costs pennies. It is also how you turn a phone photograph into collage material.

One warning you will meet properly in Lesson 16

Everything you cut out came from somewhere. A magazine advertisement, a newspaper photograph, a printed pattern: each was made by somebody who owns it. Collage has a long tradition of using such material, and the law around it is genuinely unsettled rather than simply permissive. For now, three working habits. Keep a list of what you cut and where it came from. Prefer material you photographed or made yourself, and material clearly out of copyright. And be suspicious of any collage whose interest comes entirely from one borrowed photograph, because that is the case where the borrowing is doing all the work. Lesson 16 gives you the actual legal picture.

Common misconceptions

• "Collage is what you do when you cannot draw." Bearden could draw at a very high level and chose scissors because a cut join between two unlike images says something a continuous drawing cannot.
• "Any interesting texture will do." Where the material came from is part of the meaning. Höch cut faces her readers recognised; that is why the montage bites.
• "Mixed media means using lots of materials." It means the materials do different jobs. Four media doing the same job is just a mess with a longer materials list.
• "Glue is glue." Glue stick lifts within months, rubber cement stains, and a corner dotted with PVA curls. Thin PVA brushed over the whole back is the default.
• "I can work on cartridge paper." Wet adhesive shrinks as it dries and buckles thin paper permanently. Use card, board or a stretched and taped sheet.

What to carry forward

Bearden built eighteen feet of Harlem out of cut paper and photographs, and put real photographic faces in windows made of advertisements, because the mismatch between the two is the point. A cut edge is a join between materials that were never together, and the viewer reads the join. Picasso started it in May 1912 by gluing machine printed oilcloth imitating chair caning into a painting, introducing an image made by someone else for another purpose into a work of art; Schwitters went on to build pictures out of tram tickets. Photomontage sharpened it into a weapon, and Hannah Höch's Berlin montage of 1919 to 1920 works because the scales are deliberately wrong and the faces were recognisable. Matisse, cutting shapes out of gouache painted sheets in 1953, got contour and colour in one action, and pinned everything up until it was right, which is collage's great advantage: the composition can be moved until five minutes before you commit. Choose your edges deliberately, use thin PVA over the whole back, work on board, and remember where every scrap came from. Next lesson picks up the other machine that makes images for you, the one in your pocket.

Sources

1. Bearden, R. (1971). The Block [Cut and torn pasted papers, foil, graphite, watercolour, gouache and ink on Masonite]. The Metropolitan Museum of Art, New York. metmuseum.org
2. Matisse, H. (1953). The Snail [Gouache on paper, cut and pasted]. Tate, London. tate.org.uk
3. Tate. Collage [Art term]. Tate, London. tate.org.uk
4. Wikipedia contributors. (2024). Still Life with Chair Caning. Wikipedia. en.wikipedia.org
5. Wikipedia contributors. (2024). Hannah Höch. Wikipedia. en.wikipedia.org

You stood in front of something worth photographing. The light was extraordinary, or the face was, or the shadow across the wall was exactly the shape you had been trying to draw for a fortnight. You lifted your phone, tapped the button, looked at the result, and it was nothing. Flat, grey, ordinary, smaller than the thing itself in every way that mattered.

That gap is not a limitation of the phone. The sensor behind that lens is smaller than your smallest fingernail and it still resolves more detail than a professional camera of 1990. The gap is between what you saw and what you told the camera to record, and almost all of it can be closed with decisions rather than equipment.

Diagnosing the flat photograph

Four causes, in the order they usually apply.

1. You exposed for the whole frame instead of for the subject. The camera averaged everything, so the bright window became mid grey and so did the face in front of it.
2. The light was flat. Overhead midday sun or a bright overcast sky produces almost no directional shadow, and shadow is what makes form read, exactly as in the value lesson.
3. You were too far away and zoomed. Digital zoom crops and enlarges pixels; it adds no information and subtracts sharpness.
4. You framed the way you stand, not the way you compose. Eye height, subject centred, everything in the frame, nothing cropped. This is precisely the dead still life from Lesson 10, in a different medium.

Fix those four and the equipment stops being the constraint. Fix none of them and no camera on earth will help.

Three controls, and the two you can actually reach

Exposure is set by three things acting together, and the classical name for them is the exposure triangle.

Control	What it changes	Side effect	On a phone
Aperture	How wide the lens opening is, so how much light per instant	Depth of field: wide open gives a shallow zone of sharpness	Usually fixed; you cannot change it
Shutter speed	How long the sensor is exposed	Motion: slow blurs movement, fast freezes it	Adjustable in pro mode or a manual camera app
ISO	How strongly the signal is amplified	Noise: high ISO gives speckled shadows	Adjustable in pro mode; the phone raises it in dim light

On a standard phone camera app you effectively have one control, and it is the important one. Tap the screen on your subject: this sets both focus and metering there. Then, without lifting your finger, slide up or down. That is exposure compensation, and it tells the camera to make the picture brighter or darker than its own average. Learning that one gesture will improve your photographs more than any other single thing in this lesson.

Why this matters: a camera meters for an average grey. If your subject is a white ceramic bowl or a very dark object, the camera will get it wrong every time in a predictable direction, and the slider is how you overrule it.

Depth of field, and why phones give you everything sharp

Depth of field is the distance between the nearest and furthest things that look acceptably sharp. It shrinks when the aperture is wide, when the focal length is long, and when the subject is close.

A phone camera has a very short actual focal length and a small sensor, which together produce enormous depth of field: nearly everything from a metre away to the horizon is in focus. That is convenient and it is also why phone photographs often lack the separation between subject and background that makes a photograph feel deliberate. Portrait mode fakes the difference in software by detecting the subject and blurring the rest, which works well on a head against a wall and produces strange results on hair, glasses, leaves and any edge that is genuinely complicated.

The reliable optical answer, which requires nothing, is to move. Get physically close to your subject and put a lot of distance between the subject and the background. Real depth of field appears, edges stay honest, and the background falls away without any software guessing where the subject ends.

Light first, subject second

Photographers who consistently produce good pictures generally choose the light before the subject.

• Direction. Front light flattens. Side light, where the sun or window is at ninety degrees to your view, produces the six zone structure you learned to draw: light, halftone, core shadow, reflected light, occlusion, cast shadow. Back light gives silhouette and rim light, and requires you to slide the exposure up or lose the subject entirely.
• Quality. A small source close to nothing, such as bare sun or a phone torch, gives hard edged shadows. A large source, such as an overcast sky or a north facing window, gives soft ones. You can convert hard to soft by putting a white sheet or a piece of baking parchment between the sun and your subject.
• Colour. Light has a colour temperature measured in kelvin. A household tungsten bulb is around 2700 K and is distinctly orange; midday daylight is around 5500 K; open shade under a blue sky is much higher and reads blue. Your camera guesses a correction, called white balance, and it guesses badly in mixed light, which is why the drawing photographed under a ceiling bulb comes out the colour of weak tea.
• Time. The hour after sunrise and the hour before sunset give low, warm, directional light with long shadows, which is why it is called the golden hour. The same subject at one in the afternoon and at seven in the evening is two different subjects.

Stieglitz, one plate, one moment

In 1907 Alfred Stieglitz was crossing to Europe in first class on the Kaiser Wilhelm II. By his own account he walked to the end of the deck, looked down at the lower deck where the steerage passengers travelled, and saw an arrangement: a round straw hat, a funnel leaning left, a gangway cutting a white diagonal, a chain, suspenders, figures stacked on two levels. He went back for his camera, returned hoping nothing had moved, and made one exposure. The Steerage exists in several printings, and the Metropolitan Museum of Art holds a photogravure of about 32 by 26 centimetres.

What he described noticing was not a scene of poverty. It was shapes: the hat, the diagonal, the division of the frame into an upper and a lower world by the gangway. This is worth sitting with, because it contains the whole difference between taking a photograph of a thing and making a photograph. The subject matter was there for anyone on that deck. The picture came from seeing an arrangement in it and having one exposure to catch it.

You have an unlimited number of exposures, which is an advantage and a trap. The trap is that free frames encourage you to photograph instead of looking. Try Stieglitz's constraint deliberately: give yourself six frames for an afternoon and you will find you start composing before you lift the phone.

Lange, and the frames you never see

In March 1936, at a pea pickers' camp in Nipomo, California, Dorothea Lange photographed a mother and her children for the federal Resettlement Administration. She made a series of exposures, moving closer with each one, and the last is the picture the world knows as Migrant Mother. The woman was later identified as Florence Owens Thompson. The negatives are held by the Library of Congress and the whole sequence can be seen.

Two lessons sit in that sequence. The first is craft: the earlier frames are wider, include a tent and more children and more context, and are ordinary. The last frame is close, cropped, and organised around one face and one hand. Getting closer was the entire editorial act.

The second is ethical, and it does not resolve neatly. Thompson said in later years that she felt exploited, that the photograph had made others famous and had not helped her. Lange was working for a government agency documenting rural poverty, and the picture changed how a country understood a crisis. Both of those are true. When you photograph a person, especially a person in difficulty, you are taking something. Ask. Say what the picture is for. Show them the result. Those three habits cost nothing and settle most of the problem before it starts.

Phone technique that actually changes the picture

• Clean the lens. It lives in a pocket. A fingerprint reduces contrast across the entire frame, and this is the single most ignored piece of advice in photography.
• Tap to set focus and exposure, then slide. Some phones let you press and hold to lock both, which stops the exposure jumping as you recompose.
• Never use digital zoom. Walk closer, or crop afterwards from the full resolution file, which gives an identical result with more control.
• Turn on the grid. It puts the rule of thirds lines on the screen and, more usefully, tells you whether your horizon is level.
• Shoot at the highest quality your phone offers, and if it offers a raw format, use it for anything you will edit hard. Raw keeps far more information in the highlights and shadows.
• Brace yourself. Elbows in, back against a wall, breathe out and then press. In dim light this is the difference between sharp and nearly sharp.
• Use the self timer even when you are not in the picture, because pressing the button is what shakes the camera.

Editing, without destroying anything

Editing is part of photography and always has been; darkroom printers dodged and burned every print. The discipline is that the original must survive.

Work on a copy, or in an application that edits non destructively by storing your adjustments as instructions rather than baking them into the pixels. Then make the small number of adjustments that actually matter: straighten, crop, set the black point and the white point so the image uses its full range, adjust exposure, correct the white balance, and only then consider anything else. Convert to black and white if the picture is about value rather than colour, and check that decision by looking at whether the colours were doing any work.

Two failure modes to avoid. Clipping, where you push contrast so far that shadows become featureless black and highlights become blank white, and information disappears permanently. And oversharpening, which produces bright halos along every edge and is instantly recognisable.

Key idea: keep the original file untouched, always. An edit you regret is recoverable; an overwritten original is not.

Photographing as a source for your own drawing

You will use photographs as reference, and you should. But take your own wherever you possibly can, for three reasons. A photograph you took contains information you remember, so you can draw what the camera missed. A photograph you took has no copyright problem at all, which the next lesson will show is not a small matter. And composing through a viewfinder is compositional practice in itself.

When you photograph for reference rather than for the picture, change what you shoot. Take the same subject from several distances. Take one deliberately overexposed frame for shadow detail and one underexposed for highlight detail. Photograph the light and the shadow separately if you must. A reference photograph is a set of notes, not an image, and it does not have to be good.

Common misconceptions

• "I need a better camera." The four causes of a flat photograph are exposure, light, distance and framing, and none of them is fixed by buying anything.
• "Zoom brings it closer." Digital zoom crops and enlarges pixels. It adds nothing that cropping afterwards would not give you, and it costs sharpness.
• "Portrait mode gives real depth of field." It estimates a subject and blurs the rest. Moving close to your subject, with the background far away, gives the real thing and does not fail on hair.
• "Editing is cheating." Every darkroom print was dodged and burned. Overwriting your original file is the actual mistake.
• "A photograph of a person is just a picture." Florence Owens Thompson felt the most famous photograph of the Depression had taken something from her. Ask, explain, and show people the result.

Pulling it together

The disappointing photograph almost always has one of four causes: the camera metered the whole frame instead of the subject, the light was flat, you zoomed instead of walking, or you framed at eye height with everything centred. Your phone gives you one meaningful control, the tap and slide that sets focus and then overrules the camera's assumption that the world averages to grey. Small sensors give huge depth of field, so separation comes from moving close with the background far away rather than from software blur. Choose light by direction, quality, colour temperature and time of day before you choose a subject. Stieglitz saw a hat, a funnel and a diagonal gangway on a crossing in 1907 and had one plate; Lange got closer with every exposure at Nipomo in 1936, and the last frame is the one the world knows. Clean the lens, use the grid, brace, shoot at full quality, edit on a copy, never clip the ends of the range, and take your own reference photographs. Next lesson takes these files and opens them in software that costs nothing.

Sources

1. Stieglitz, A. (1907). The Steerage [Photogravure]. The Metropolitan Museum of Art, New York. metmuseum.org
2. Wikipedia contributors. (2024). Migrant Mother. Wikipedia. en.wikipedia.org
3. Wikipedia contributors. (2024). Exposure (photography). Wikipedia. en.wikipedia.org
4. Wikipedia contributors. (2024). Depth of field. Wikipedia. en.wikipedia.org
5. Tate. Photography [Art term]. Tate, London. tate.org.uk

In 1995 two students at the University of California, Berkeley, Spencer Kimball and Peter Mattis, began writing an image editor as a semester project in the eXperimental Computing Facility. They released version 0.54 publicly in 1996 and called it the General Image Manipulation Program. Richard Stallman visited Berkeley in 1997, met them, and agreed that the G could stand for GNU instead. GIMP has been free, in both senses, ever since.

That matters to you for a practical reason. The tools a professional illustrator or photographer uses are available to you at no cost, on any computer, today. What is not free is knowing which tool to open and what to set before you make the first mark, and getting those wrong is why so much student digital work looks thin.

Raster and vector, the distinction everything else hangs on

	Raster	Vector
What the file contains	A grid of coloured pixels	Mathematical descriptions of points, curves and fills
What happens when you enlarge it	The pixels get bigger; detail is invented or blocky	Nothing; it redraws perfectly at any size
What it is good at	Photographs, painting, texture, anything continuous	Logos, type, diagrams, flat shapes, anything to be printed at several sizes
What it is bad at	Being enlarged, and being edited after the fact	Soft edges, photographic texture, subtle brushwork
Typical files	PNG, JPEG, TIFF, XCF, KRA	SVG, PDF, AI, EPS
Free tool	GIMP, Krita	Inkscape

Almost every frustration in student digital work traces back to using one where the other was needed. A logo drawn in a raster program is a logo that cannot be printed on a poster. A photograph traced into vector paths is a photograph that has lost everything that made it a photograph.

The point: raster stores what the image looks like; vector stores how to draw it. Choose according to whether the image needs to survive being resized.

Resolution, and the number to set before you start

A raster image has a fixed number of pixels and no more. Enlarging it later cannot add information, and every upscaling algorithm is an educated guess.

The working number for print is 300 pixels per inch at final size. To print at A4, which is roughly 8.3 by 11.7 inches, you therefore need an image of about 2480 by 3508 pixels. If you begin a digital painting on a canvas of 1000 by 1400 pixels because it feels responsive, you have already decided that the piece can never be printed larger than a postcard at full quality, and you will discover that on the evening the portfolio is due.

So set the dimensions first. Decide the largest size the piece could ever need to be printed at, multiply by 300, and start there. If your machine struggles, work at that size and reduce the brush complexity, not the canvas.

Screen colour is not paper colour

A screen makes colour by adding red, green and blue light: black is the screen off, white is all three at full. Printing makes colour by subtracting, laying cyan, magenta, yellow and black ink over white paper so that each layer removes some of the light bouncing back. These are opposite systems, and they do not cover the same set of colours.

The practical consequence is that some colours you can see on your screen, particularly saturated oranges, bright greens and electric blues, cannot be printed at all. They will come out duller. This is not a fault in your printer; it is what a gamut is. If a piece will be printed, check it in a soft proof if your software offers one, and expect the most saturated colours to lose intensity.

Work in sRGB unless you have a specific reason not to. It is the standard colour space of the web and of almost every screen your work will be seen on, and files sent out in other spaces frequently arrive looking wrong.

Layers and masks, or: stop using the eraser

A layer is a transparent sheet stacked over the ones below. Every serious digital tool is built on them, and using them properly is the difference between digital work you can revise and digital work you have to restart.

Three habits change everything.

1. Put separable things on separate layers. Line work on one, flat colour beneath it, shadows above, background at the bottom. When you decide at midnight that the background should be warmer, it is thirty seconds of work rather than an hour of careful selection.
2. Use a layer mask instead of the eraser. A mask is a greyscale channel attached to a layer: paint black on the mask and that part of the layer becomes invisible, paint white and it returns. Nothing is destroyed, so you can change your mind about an edge indefinitely. An erased pixel is gone the moment you save and close.
3. Name your layers. A file with thirty layers called Layer 12 copy copy is unusable a week later, and a week later is exactly when you will need it.

Blend modes are the other thing worth learning, and only three of them matter at first. Multiply darkens and is how you lay shadow over flat colour, because it behaves like a transparent glaze. Screen lightens. Overlay increases contrast. Everything else is worth exploring later.

Four programs that cost nothing

Tool	Use it for	Notes
GIMP	General raster editing, photo retouching, compositing, collage from scans	The oldest and broadest; native format XCF
Krita	Digital painting and drawing	Built by and for painters; the brush engines and stabiliser are far better suited to drawing than a general editor
Inkscape	Vector work: type, diagrams, flat shape design, anything scalable	Saves plain SVG, which opens everywhere
darktable	Developing raw photographs and processing a whole shoot consistently	Fully non destructive; your raw files are never altered

Blender, also free, does three dimensional modelling and rendering, and is worth knowing exists if your investigation goes that way. A graphics tablet transforms drawing in Krita, because pressure sensitivity restores the varied line weight you worked on in Module 2, but none of this requires one.

A workflow that produces something you can revise

Take a concrete case: you have a pencil drawing on paper and you want a coloured digital version that still looks like your drawing.

1. Photograph or scan the drawing at high resolution, flat and evenly lit, following the next lesson's method.
2. Open it in GIMP or Krita and immediately save as the native format, XCF or KRA, under a new name. The original file stays untouched.
3. Straighten and crop, then set the black and white points so the paper is white and the darkest pencil is nearly black.
4. Put the line work on its own layer and set that layer to Multiply, so the white of the paper becomes transparent and only the pencil shows.
5. Add a new layer beneath it for flat colour, and paint underneath the drawing. Your pencil lines stay on top and stay visible.
6. Add shading on a layer above the flats, also set to Multiply, and use a mask rather than an eraser to shape it.
7. Export a copy as PNG for showing, and keep the layered file. Never overwrite the layered file with the flattened one.

The file format that quietly destroys your work

JPEG is a lossy format: it discards information to make the file small, and it discards more every time you save. Open a JPEG, adjust it, save it, open it again next week, adjust it, save it again, and you are compressing a compression. After a dozen rounds the edges carry visible blocky artefacts that nothing can remove.

• Working files: the native format of your program, XCF, KRA or SVG, with layers intact.
• Lossless export: PNG, which also supports transparency, or TIFF for print.
• JPEG: for sending and uploading only, exported fresh each time from the layered original.
• Backups: at least two copies in different places. A portfolio that exists only on one laptop is one spilled drink away from not existing.

Name files so they sort and so you can read them a year later. Something like investigation-03-window-study-2026-04-12-v2 tells you everything; something like final-final-2 tells you nothing.

So what?: the most common way students lose a term of digital work is not a dramatic crash. It is repeatedly saving over a JPEG, or flattening and overwriting the only layered file.

What the software will not do

Digital tools remove consequences, and that is both their gift and their trap. Undo, layers and masks mean you can try anything, which is genuinely liberating and makes them excellent for experimenting with composition and colour. The same absence of consequence removes the pressure that produces decisions. A charcoal drawing forces you to commit; a digital file lets you defer forever, and a piece can be nudged for eleven hours without a single decision being made.

Set yourself artificial constraints to get it back. Give a digital study a time limit. Work at a fixed number of layers. Disallow undo for the first twenty minutes. And when a digital piece stalls, print it, draw on the print in pen, and photograph that back in. Moving between media is usually faster than pushing harder in one.

Common misconceptions

• "I can enlarge it later." A raster image has a fixed number of pixels and enlargement invents the rest. Set the final dimensions before you start.
• "The colours will print as they look." Screens add light and printers subtract it. The most saturated screen colours sit outside a printer's gamut and will come back duller.
• "The eraser is the digital rubber." It destroys pixels. A layer mask hides them and lets you change your mind next week.
• "Free software is a compromise." GIMP dates from 1995 and Krita's painting tools are built specifically for artists. Both are used professionally.
• "JPEG is just a picture file." It is lossy and it degrades every time you resave. Keep the layered original and export a fresh JPEG when you need one.

Where this leaves us

Two students at Berkeley wrote GIMP as a semester project in 1995 and released it in 1996, which is why the tools of professional image making cost nothing today. Raster stores what an image looks like and vector stores how to draw it, so choose by whether the piece must survive resizing. Set your dimensions at 300 pixels per inch at final print size before the first mark, because enlargement invents information. Screen colour is additive and print colour subtractive, so expect saturated colours to dull, and work in sRGB. Put separable things on separate layers, name them, mask rather than erase, and learn Multiply first. GIMP, Krita, Inkscape and darktable cover raster, painting, vector and photographs between them. Keep layered originals in the native format, export PNG for lossless copies and JPEG only for sending, and back up in two places. Above all, notice that removing consequences also removes the moment of decision, and impose a constraint to put it back. Module 6 turns to everything around the work: where your sources came from, how you photograph and describe what you made, and how eighteen pieces become one portfolio.

Sources

1. GIMP Development Team. GNU Image Manipulation Program: official site and documentation. GIMP. gimp.org
2. Krita Foundation. Krita: free and open source digital painting application. Krita. krita.org
3. Inkscape Project. Inkscape: free and open source vector graphics editor. Inkscape. inkscape.org
4. Wikipedia contributors. (2024). GIMP. Wikipedia. en.wikipedia.org
5. Wikipedia contributors. (2024). Vector graphics. Wikipedia. en.wikipedia.org

In 1981 the photographer Lynn Goldsmith photographed the musician Prince on assignment for Newsweek. The picture was not published, and she kept the copyright. Three years later Vanity Fair licensed that photograph so that Andy Warhol could use it as the reference for a silkscreen illustration, on the agreed terms that it be published once and that Goldsmith be credited. Warhol went on to make a whole series of Prince images from it, and Goldsmith did not know. She found out in 2016, after Prince died, when Condé Nast put one of them, Orange Prince, on the cover of a tribute magazine with no credit to her, under a licence from the Andy Warhol Foundation.

The litigation ran seven years. A district court sided with the Foundation in 2019, the Second Circuit reversed in 2021, and in 2023 the Supreme Court affirmed the reversal. That case is the sharpest available test of a question you will face this year the first time you open a search engine looking for something to draw: what exactly may you take, and from whom?

This lesson is not legal advice, and it describes United States law, which is not the law everywhere. It is enough to keep you out of trouble and, more importantly, to let you think clearly about what you owe a source.

Two positions, each with real weight behind it

Set the strongest version of each argument next to the other, because the honest answer is that they have not been reconciled.

The artist's position. Images now circulate in enormous quantities and are part of the shared visual environment, in the way that folk songs once were. Art has always reworked existing material: Manet reworked Titian, Rauschenberg silkscreened press photographs, Bearden cut faces from magazines. Rules that forbid this would have made much of twentieth century art illegal, and they tend to favour whoever owns the most images rather than whoever has something to say. Richard Prince's Canal Zone works, made in 2008 from Patrick Cariou's photographs, were held in 2013 by the Second Circuit to be, in a number of instances, transformative fair use.

The photographer's position. A photograph is not raw material lying about. Somebody chose the subject, waited for the light, framed it, exposed for it and printed it, and every one of those decisions is exactly the kind of decision this course has spent five modules teaching you to make. Licensing that work is how most photographers eat. Lynn Goldsmith is a working photographer whose photograph was used, for money, on a magazine cover without payment or credit. Art Rogers photographed a couple holding puppies for a greeting card; Jeff Koons found it on a postcard, removed the copyright notice, and had assistants build a sculpture from it that sold in three copies for a total of $367,000.

Both positions are held sincerely by serious people. What follows is not a resolution. It is the framework the courts actually use, and then a working rule for a student portfolio.

What copyright protects, briefly

• It protects original expression fixed in a tangible form: this photograph, this drawing, this recording. It arises automatically the moment the work is made, with no registration and no notice required.
• It does not protect ideas, facts, subjects, techniques or styles. Nobody owns still lifes of lemons, the idea of photographing a bridge in the rain, or drawing in crosshatch.
• For works made from 1978 onwards in the United States it lasts for the life of the author plus seventy years.
• Older works fall out of protection on a rolling schedule. As things stand, works published in the United States in 1930 or earlier are in the public domain, and that line moves forward by one year every first of January.
• It is separate from the ownership of the physical object. Buying a print does not buy the right to reproduce it, and being in a photograph does not make you its copyright owner.

That third and fourth point together explain why every artwork named in this course as a source is either very old or linked to a museum that has chosen to publish an image of it.

The four factors, and what actually moves them

Fair use is a defence in United States law, decided case by case against four statutory factors. The Copyright Office maintains a searchable index of how courts have applied them.

Factor	The question	What helps you	What hurts you
1. Purpose and character	Why are you using it, and does your use add a different meaning or purpose?	Commentary, criticism, parody, teaching, a genuinely different function	A commercial use that serves the same purpose as the original
2. Nature of the work	What kind of thing did you copy?	Factual material, already published work	Highly creative work, and unpublished work
3. Amount and substantiality	How much did you take, and was it the heart of the work?	A small part, no more than your purpose needs	The whole thing, or the one element that made it what it was
4. Market effect	Does your use substitute for the original or damage its licensing market?	No overlap with where the original earns	Competing with the original, or with the licences its owner would sell

Two things students consistently get wrong here. Changing a fixed percentage of an image is not a rule and never has been; there is no ten percent, no thirty percent, no seven changes. And crediting the source is good manners and good scholarship, but it is not permission, and it is not a defence.

Where copying clearly failed

Rogers v. Koons, decided by the Second Circuit in 1992, is the cleanest example. Koons wanted a sculpture for a show about banality, took Rogers's photograph from a postcard, and instructed his fabricators to reproduce as much detail as possible, with the puppies blue, the noses exaggerated and flowers added to the couple's hair. The court found the copying obvious to an ordinary viewer and rejected the parody argument, on the reasoning that Koons had no need to imitate that particular photograph to comment on banality in general. If your commentary is about the world rather than about this specific image, you did not need this specific image.

Where copying won, and settled nothing

Patrick Cariou published Yes, Rasta in 2000, a book of black and white photographs of a Rastafarian community in Jamaica. Eight years later Richard Prince made Canal Zone, a series that printed Cariou's photographs, enlarged them, blurred or sharpened them, painted over them and combined them with other material. In 2013 the Second Circuit held that a number of the works were transformative fair use as a matter of law and sent five back for reconsideration; the Supreme Court declined to hear it and the case settled in 2014.

Cariou was criticised almost immediately, by academics and by other courts, for putting too much weight on how different the new work looked. The trouble with a test that asks whether a work has a new aesthetic is that it invites judges to decide which art is sufficiently new, which they are not well placed to do and which produces unpredictable results for everyone.

What 2023 changed

In the Warhol case the Supreme Court narrowed the first factor considerably. The question, it held, is not whether the artist added something or had a new artistic purpose in the abstract. It is what the particular use being challenged was for. The specific use at issue was the licensing of Orange Prince to a magazine to illustrate an article about Prince, which is exactly what Goldsmith's photograph would have been licensed for. Same purpose, commercial, in her market. Factor one favoured Goldsmith.

Read carefully, that decision does not outlaw appropriation and it did not rule on the Prince Series hanging in a museum. It says that the analysis attaches to the use rather than to the artist's reputation or intention. The same image can be fair use in one context and infringing in another.

Bottom line: the question is never whether you changed it enough. The question is what your version is for, and whether it lands in the same place the original earns its living.

Reference against appropriation, as a working distinction

Legal categories are not studio categories. Here is the studio version, which is stricter than the law and will keep your portfolio clean.

	Reference	Appropriation
What you take	Information: how a wrist bends, what a rusted hinge looks like, the shape of a wave	The image itself, recognisably, including its composition and framing
Who made the decisions	You: your composition, lighting, crop, format	The original author, and you are keeping them
Would a viewer recognise the source?	No; they would see your subject	Yes, and usually that recognition is the point
What it needs	Nothing, though your own photographs are still better	A reason, an acknowledgement, and a hard look at the four factors

The hardest case is the honest one: you find a photograph on the internet of exactly the thing you want to draw, in exactly the light you want, and you draw it faithfully. That drawing is not reference. Everything that makes the image good, the angle, the light, the crop, the moment, was decided by the photographer, and your contribution is rendering. It is also, in United States law, likely to be a derivative work of theirs.

Why it is a problem even when it is legal

Set the law aside for a moment. Five modules of this course have argued that the decisions are the work: which crop, which light, which value structure, which moment. When you draw someone else's photograph faithfully, you have accepted every one of those decisions ready made and spent your afternoon on execution alone.

That has a practical cost as well as an ethical one. Copying photographs is a well known way to plateau. You get very good at rendering and never get better at seeing, because the seeing was done for you. Anyone assessing a portfolio can spot it: the drawings have the flattened value structure, the frozen motion and the characteristic lens distortion of a photograph, and the sketchbook contains no evidence of the subject having been observed. If you must work from a photograph, work from one you took.

Images you can use without worrying

• Your own photographs. Free, unencumbered, and full of information you remember. Always the first answer.
• Public domain works. Older published works, on the rolling schedule above. A Dürer engraving of 1514 is nobody's property.
• Museum open access programmes. Several major museums publish images of public domain works in their collections for unrestricted use. Check the individual object page, because the policy is per object rather than per museum.
• Creative Commons licensed images. These are copyrighted works whose owner has granted permission in advance, on conditions. Read which licence: some require attribution, some forbid commercial use, some require that your resulting work carry the same licence.
• Asking. Photographers are people with email addresses. A polite message from a student explaining the project is answered far more often than you would expect, and a yes in writing ends the problem entirely.

The upshot: the supply of images you can use freely is enormous, and reaching for a random search result is a choice, not a necessity.

A source log that takes thirty seconds a time

Keep a running list at the back of your sketchbook or in one file. For every image you did not make: what it is, who made it if you can find out, where you got it, the date, and one line on how you used it. Write it when you save the file, not at the end of the year, because at the end of the year you will not remember.

This does three jobs. It answers the question if anyone asks it. It stops you from accidentally building a finished piece on a source you cannot account for. And it turns into the evidence of research that a portfolio is supposed to contain anyway, which means the log is not extra work but a part of the work you were already meant to be doing.

Common misconceptions

• "If I change thirty percent it is mine." There is no percentage rule and never has been. Courts weigh four factors, and the first one asks what your use is for.
• "If I credit the photographer it is fine." Attribution is good practice and is not permission. Condé Nast's use of Orange Prince carried no credit, but credit would not have decided the case.
• "It is on the internet, so it is public." Copyright attaches automatically at creation. Being findable has nothing to do with being free.
• "It is for school, so it is fair use." Educational purpose helps under factor one but does not decide it, and portfolios get published, exhibited and sold.
• "Copying photographs is how you learn to draw." It is how you learn to render. The seeing, which is the harder half, was done by the photographer.

Putting it together

Lynn Goldsmith photographed Prince in 1981, licensed the picture once to Vanity Fair as an artist reference, and learned only in 2016 that a whole series had come out of it and that one of them was on a magazine cover without her name. Copyright protects fixed original expression automatically, not ideas, subjects or styles, and it runs for the author's life plus seventy years, with older published work falling into the public domain on a rolling schedule. Fair use is decided on four factors, and there is no percentage rule and no credit exemption. Koons lost in 1992 because he did not need that particular photograph to comment on banality; Richard Prince largely won in 2013 on a test that asked how different the work looked, and that test was criticised for asking judges to grade art. The Supreme Court in 2023 shifted the question to what the specific use is for, so the same image can be fair in one context and infringing in another. In the studio, reference means taking information and making the decisions yourself, while appropriation means keeping the author's decisions, and the second needs a reason. Use your own photographs, public domain works, museum open access images and Creative Commons material, keep a source log as you go, and remember that copying photographs teaches rendering while quietly costing you the ability to see. Next lesson: how to photograph and describe the work you legitimately made.

Sources

1. U.S. Copyright Office. What is Copyright? Library of Congress. copyright.gov
2. U.S. Copyright Office. Fair Use Index. Library of Congress. copyright.gov
3. Supreme Court of the United States. (2023). Andy Warhol Foundation for the Visual Arts, Inc. v. Goldsmith, 598 U.S. 508 [Opinion of the Court]. supremecourt.gov
4. Wikipedia contributors. (2024). Rogers v. Koons. Wikipedia. en.wikipedia.org
5. Wikipedia contributors. (2024). Cariou v. Prince. Wikipedia. en.wikipedia.org

It is half past ten the night before the deadline. A charcoal drawing is lying on a kitchen table under a single ceiling bulb, and a phone is being held above it at an angle, because holding it straight overhead puts the photographer's head in the way. The resulting file shows a drawing that is orange, trapezoidal, brighter at the top than the bottom, with a shadow of two hands across one corner and a fruit bowl in the frame.

That file is now the drawing, as far as anyone else is concerned. Nobody who assesses your portfolio will see the paper. This lesson takes that photograph apart fault by fault, then does the same to the paragraph that usually accompanies it.

Five faults, and where each one came from

1. The trapezoid. The camera was not square to the work, so the near edge is wider than the far one. This is keystone distortion, and it is the fault that most immediately says amateur, because a rectangle that is not a rectangle is instantly visible.
2. The orange cast. A household tungsten bulb sits near 2700 K, and the camera's automatic white balance guessed wrongly. The paper is not white in the file, and every value in the drawing is therefore reported wrongly.
3. The uneven light. One bulb overhead means the part of the paper nearest the bulb is brighter. A drawing whose value structure you spent six hours on is now being read through a gradient you did not draw.
4. The shadow. The light source was behind the photographer, so the photographer is in the picture. Any light directly behind you will do this.
5. The fruit bowl. Nothing that is not the work belongs in the frame.

The fix, in order, with nothing you do not own

1. Find even light. The easiest good light in the world is free: outdoors on an overcast day, or just inside a north facing window with the direct sun off the work. Failing that, two identical lamps, one on each side, at roughly forty five degrees to the surface and at the same distance, which cancels each other's gradient. Turn the ceiling light off; mixed light sources are what defeat white balance entirely.
2. Get square. Put the work flat on the floor and stand directly over it, or tape it to a wall and put the camera on a stack of books at the height of the work's centre. Use the phone's grid and level. If the edges of the paper are parallel to the edges of the frame, you are square.
3. Fill the frame, then leave a margin. Get close enough that the work occupies most of the frame, with a small even border of neutral background. Do not use digital zoom; move.
4. Fix the white balance. Photograph a plain sheet of white paper in the same light and use it to set white balance in editing, or shoot raw and correct afterwards. A grey card or a colour reference chart does this properly and costs very little, but a blank sheet of your own paper works well.
5. Include a reference frame, outside the final crop. Lay a ruler along one edge, and if you have one, a colour reference. Crop both out of the version you submit and keep the uncropped file. You will one day need to know exactly how big something was and exactly what colour.
6. Check the corners. Before you put the camera down, look at all four corners of the frame for a hand, a shadow, a cable or a fruit bowl.
7. Edit lightly. Straighten, crop to the work's edges, correct white balance, set the black and white points so the paper reads as paper and the darkest mark reads as dark. Stop there. Editing a photograph of an artwork until it looks better than the artwork is a form of lying, and it is obvious to anyone who then sees the real thing.

Remember: the aim is a file that a stranger could use to reconstruct your decisions. Not a flattering picture, an accurate one.

Three cases that need something extra

Case	The problem	What to do
Glossy or varnished surfaces, and work behind glass	Specular reflections of the lights and of you	Take the glass off if you can. Otherwise light from a steep angle to each side, wear dark clothes, and stand back with the camera further away
Three-dimensional work	A single frontal shot flattens it and hides how it occupies space	Shoot four to six angles on a plain uncluttered background, with one raking light to reveal form and a soft fill on the other side, plus at least one shot with something for scale
Large work, and sketchbook spreads	You cannot get far enough back, or the book will not lie flat	Photograph large work outdoors on an overcast day. For a sketchbook, weight the pages open, shoot from directly above, and photograph the spread rather than each page, since the relationship between the pages is part of the evidence

What else to record, and when

The finished photograph is the smallest part of your documentation. What a portfolio is short of is almost never finished work; it is evidence of the thinking that produced it.

• Process photographs. One at the block in stage, one at the halfway point, one before the last change. Thirty seconds each. Take them even when the piece is going badly, especially then.
• Details. One or two close photographs of an area where the mark making does something worth seeing. A full view at portfolio resolution loses the surface entirely.
• Scale. Either a ruler in an uncropped version or a written dimension. A viewer looking at a screen has no idea whether your drawing is A5 or two metres, and the difference changes everything about it.
• Dates. Written on the back of the work in pencil, and in the file name. Chronology is what turns a group of pieces into an investigation.
• The failures. As Lesson 1 argued, the abandoned version and the change you made in response to it are the evidence that you were deciding rather than habitually producing.

Do this weekly. Documentation collected in one panicked evening in May is thin, and it is thin in a way that is visible.

The statement that says nothing

Here is a statement of the kind that arrives by the thousand. It is grammatical, it is sincere, and it is empty.

My work explores themes of identity and memory. I am interested in how we perceive the world around us and how our experiences shape who we are. Through my use of colour and texture, I invite the viewer to reflect on their own journey. I have always been passionate about art, and I hope my work resonates with others.

Take it apart line by line. Sentence one names two abstractions, identity and memory, that could sit in front of almost any body of work made in the last century. Sentence two is a claim about human beings in general, not about anything you made. Sentence three describes materials at a level so general that it excludes nothing: all painting uses colour, all surfaces have texture. It also hands the work's job to the viewer. Sentence four is a biography of your enthusiasm and a wish.

Now the test from Lesson 10, applied to writing: could this paragraph sit unchanged in front of somebody else's portfolio? It could sit in front of most of them. That is the whole diagnosis.

Rebuilding it

A statement that works answers five questions, in concrete terms, in this order.

1. What did you actually make? Number, medium, size, over what period.
2. What question were you asking? Stated as a question, in your own words, not as a theme.
3. What did you do to answer it? The method: what you held constant, what you changed, what you tried.
4. What happened? Including what did not work, and what you changed as a result.
5. Where has it got to? What the work now suggests, and what you would do next.

Suppose the investigation was about the light in a single room. Here is the same student, rewritten:

These fourteen pieces are drawings and monotypes of one room, the back bedroom of my grandmother's flat, made between September and April. I started with a question I could not answer by looking: how much of a room can you remove before it stops being that room? I fixed the viewpoint, always from the doorway, and changed one thing at a time. First I removed colour, then furniture, then the walls, keeping only the shapes the light made on the floor. The furniture drawings were the weakest, because I was still drawing objects rather than light, so from December I worked only from photographs taken at four in the afternoon, when the light comes through the window at a low angle. The monotypes came out of that: I could wipe ink away, which is closer to how light actually removes detail than drawing is. The room is still recognisable in the last five pieces and I no longer know why, which is what I would work on next.

Every sentence in that version contains something only this student could have written: a room, a doorway, a date, a decision, a failure, and an unresolved question. It is not more sophisticated than the first version. It is more specific, which is a different thing and the only thing that matters here.

What matters here: a statement is not a defence of the work or a description of your feelings about art. It is a report of what you did and what happened.

Length, tense, and the words to cut

• Length: usually one hundred to three hundred words unless you are told otherwise. Written responses attached to portfolio submissions are often much shorter, sometimes a hundred words or fewer, and short forces specificity.
• Tense: past for what you did, present for what the work does now. Avoid the future, which is where vague intentions live.
• Person: first person, plainly. Not the third person voice of a catalogue essay about yourself.
• Cut on sight: explores, themes of, seeks to, invites the viewer, journey, resonate, juxtaposition, dialogue, interplay, my passion for art, and the phrase in today's society.
• Read it aloud. If you would not say it to somebody standing next to you in front of the work, do not write it.

Labelling a single piece

The museum convention, sometimes called tombstone data, is worth adopting because it is complete and takes one line: artist, title, date, medium, dimensions as height then width then depth, and where it is. Look at any of the museum citations in this course and you will see the same order.

For your own work: give the piece a title that names something rather than gesturing. Untitled is a decision and usually not yours yet. Back Bedroom, Four O'Clock tells a viewer where to look. Record the medium honestly and completely: charcoal and white chalk on grey Ingres paper, not mixed media, which tells nobody anything.

Common misconceptions

• "Documentation is what you do at the end." Process photographs cannot be taken retrospectively. The stages you did not photograph are gone.
• "A good photograph of the work should make it look its best." It should make it look like itself. Editing a photograph past the work is obvious the moment anyone sees the original.
• "The statement should sound like art writing." Art writing that says nothing is the failure mode. Say what you made, what you asked, what you did and what happened.
• "Mixed media is an adequate medium line." It withholds information. List what is actually in it and on what.
• "Scale is obvious from the picture." On a screen every work is the same size. Give a dimension or put a ruler in the uncropped file.

The takeaway

The photograph is the work, for everyone who is not in the room with it. The kitchen table version fails five ways at once, and each has a free fix: even light from an overcast sky or two matched lamps at forty five degrees, the camera square so the edges of the paper run parallel to the edges of the frame, the frame filled by moving rather than zooming, white balance set from a sheet of your own paper, a ruler included outside the final crop, and light editing that stops at accurate. Glass comes off, three-dimensional work needs several angles and a raking light, and sketchbooks are photographed as spreads because the relationship between the pages is the evidence. Record process, details, scale, dates and failures weekly, because a portfolio is usually short of thinking rather than short of finished pieces. Then write a statement that answers five questions in order: what you made, what you asked, what you did, what happened including what failed, and where it has got to. Cut every sentence that could sit in front of somebody else's portfolio. Label the work the way a museum would. One lesson remains, and it is the one that needs other people: how to run a critique, and how to build fifteen pieces into a single argument.

Sources

1. Wikipedia contributors. (2024). Artist's statement. Wikipedia. en.wikipedia.org
2. Wikipedia contributors. (2024). Color balance. Wikipedia. en.wikipedia.org
3. Wikipedia contributors. (2024). ColorChecker. Wikipedia. en.wikipedia.org
4. Wikipedia contributors. (2024). Image resolution. Wikipedia. en.wikipedia.org
5. College Board. AP Drawing: about the course and the portfolio. AP Students. apstudents.collegeboard.org

Liz Lerman founded a dance company in Washington in 1976 and ran it for thirty five years. Somewhere in those decades of showing unfinished work to rooms full of people, she got tired of what happens in an ordinary feedback session and built a replacement: a four step method she called the Critical Response Process, in which the person who made the thing asks the first questions and nobody is allowed to offer an opinion until the maker has asked for it.

It was invented for choreography and it works unchanged for a drawing on a wall. This lesson gives you that protocol, a second one for when there is nobody around, and then the last job of the year, which is turning a pile of work into a portfolio that argues.

Why ordinary feedback fails

Put a drawing on a wall, ask a group what they think, and three things reliably happen. Someone begins with a judgement, usually kindly meant and usually about whether they like it. Someone else describes what they would have done, which is a description of their own work rather than yours. And you spend the whole session explaining what you intended, which means you are talking instead of listening and are defending a piece nobody has attacked.

None of that is anyone's fault. It is what an unstructured room does. The fix is not better manners; it is an order of operations that makes the useful contributions arrive before the ones that shut the conversation down.

The Critical Response Process, run properly

You need the work, the maker, at least one responder, and someone to keep the order. Twenty minutes per piece is plenty.

Step	Who speaks	What is said
1. Statements of meaning	Responders	What was meaningful, surprising, evocative or memorable about the work. Not praise, and not a verdict: an honest report of what landed
2. Artist as questioner	Maker asks, responders answer	The maker asks whatever they want to know. Responders answer, and may state opinions here because they were asked for
3. Neutral questions	Responders ask, maker answers	Questions with no opinion smuggled inside them. This is the hardest step and the most valuable
4. Opinion time	Responders, with permission	A responder names the subject of their opinion and asks whether the maker wants to hear it. The maker may say no

Step three deserves attention, because almost every question people ask is actually an opinion wearing a question mark. Compare these.

• Loaded: Why did you make the background so busy? Neutral: What were you thinking about when you decided how much detail to put in the background?
• Loaded: Have you thought about cropping it? Neutral: How did you decide where the edges would fall?
• Loaded: Don't you think the blue is too strong? Neutral: What is the relationship you want between the blue and the grey?

The loaded versions all deliver a verdict and ask the maker to accept or defend it. The neutral versions ask about a decision, and the answer usually reveals to the maker whether they made one.

Key idea: the point of a critique is not to collect verdicts. It is to find out what a viewer actually experienced, and where your intention and their experience diverge.

A protocol for when you are alone

Most weeks you will not have a room. Use the standard four stage sequence taught in art classrooms, out loud or in writing, and force yourself to complete each stage before starting the next.

1. Describe. Only what is physically there. A vertical rectangle, three quarters of it dark, a pale diagonal band from lower left, two small warm marks near the top right. No interpretation and no value judgements. Most people cannot do a minute of this without slipping.
2. Analyse. How the parts relate. Where the eye enters, what dominates, where the contrast is highest, what is repeated, how the value structure divides the rectangle. This is Lesson 10 applied to your own work.
3. Interpret. What it might mean, given only the first two stages. What does putting three quarters of the picture in darkness do to a viewer who knows nothing about your intentions?
4. Judge. Only now. Does it succeed at what stages two and three suggest it is doing? Not: do I like it.

The discipline is the sequence. Judging first, which is what everybody does, means the judgement is made on a glance and the rest of the analysis is spent justifying it.

Three cheap mechanical tricks belong here too, and all of them work by making the familiar unfamiliar: turn the work upside down, look at it in a mirror, and photograph it and look at the photograph small. Every one of them shows you the arrangement instead of the subject, which is why they have appeared in three separate lessons of this course.

Receiving a critique without defending

Four habits, and they are hard.

• Do not explain until asked. If you have to narrate your intention before a viewer can see it, that is itself the finding, and you have just contaminated the sample.
• Write, do not argue. Take notes. A pen in your hand converts the urge to reply into the act of recording.
• Separate the observation from the prescription. When someone says the top right corner is dead and you should put something there, the observation may be right and the prescription almost always is not. Keep the first, discard the second, and find your own fix.
• Count repeats. One person noticing something is data. Four people independently noticing the same thing is a fact about the work, whatever you intended.

Then decide. You are not obliged to act on feedback; you are obliged to hear it. Sort every note into act on now, note for the next piece, and disagree, and write one line on why for anything in the third pile. That last line is worth more to a portfolio than the change would have been, because it is a decision with a reason attached.

Selecting: which work goes in

You have perhaps thirty pieces from the year and room for fifteen images in a sustained investigation section plus a smaller set of selected works. Selection is where portfolios are won and lost, and the instinct that ruins them is including your favourite pieces.

Ask instead which pieces the investigation needs. A portfolio is an argument, and an argument needs premises, not highlights.

• Keep anything that shows a decision changing. The last drawing before you changed method and the first one after it are worth more together than either is alone.
• Keep a failure with its consequence. Not failures on their own, which look like weakness, but a failure next to what you did about it, which looks like thinking.
• Cut the strong piece that belongs to nothing. That accomplished portrait from October, unconnected to the question, weakens the whole by making the investigation look accidental. Put it in selected works if it fits, or leave it out.
• Cut duplicates. If two pieces make the same move, keep the better one and use the space for a different move.
• Keep the process pages. Sketchbook spreads, tests, state proofs from the printmaking lesson, rejected collage arrangements, the notan studies. These are the evidence.

Sequencing: making the images argue

Order carries meaning, which is the whole reason series exist. Monet's Rouen Cathedral canvases at the Metropolitan mean something as a group that no single canvas means. Kollwitz issued her War woodcuts as a portfolio in a fixed order rather than as seven separate prints.

Four orderings that work, and one that does not.

1. Chronological. The default, and honest: it shows development directly. Its risk is that development is rarely a straight line, and a strictly dated order can hide the actual structure of your thinking.
2. By variable. Group by what you were changing: all the pieces where you changed scale, then all the ones where you changed light. This makes the method visible and suits an investigation with a clear set of experiments.
3. Question and answer. Pair pieces: an attempt, then the response to it. Reads well and demonstrates iteration in the clearest possible way.
4. Visual rhythm. Sequence so that consecutive images differ in scale, value key or density, so a viewer moving through does not fall asleep in a run of eight similar greys.

The one that does not work is best first. Leading with your strongest piece makes everything after it a decline, and portfolios are read to the end. Open with a piece that raises the question clearly, put the strongest work at the two thirds point, and end with the piece that opens the next question rather than the one that closes this one.

The core of it: a portfolio is not a collection of your best work. It is the shortest sequence of images that makes your investigation visible and convincing.

The last two weeks

• Lay it all out physically. Print every image small, put them on a table or a floor, and move them until the order is right. You cannot judge a sequence by scrolling.
• Check the documentation. Every image straight, square, colour correct, cropped to the work. Reshoot anything that fails Lesson 17's checklist. A weak photograph of a strong piece is a weak piece.
• Read the written responses aloud. To another person, and watch their face. The sentence where they look confused is the sentence to rewrite.
• Do the source audit. Anything built on a borrowed image should be accounted for in your log, and anything you cannot account for should not be in there.
• Run the specificity test on everything you have written. Delete any sentence that could sit in front of somebody else's portfolio.
• Back it up in two places. Then check that the backup opens.

Common misconceptions

• "A critique is where people tell me what is wrong." It is where you find out what a viewer experienced. Opinions come last, by permission, and only after the neutral questions have done their work.
• "I should explain my intentions first." If the work needs narration to be seen, that is the finding. Explain afterwards, if asked.
• "Feedback must be acted on." It must be heard. Separate the observation from the prescription, keep the first, and write down why when you disagree.
• "The portfolio should contain my best work." It should contain the work the investigation needs. A brilliant unconnected piece makes the whole look accidental.
• "Put the strongest piece first." Then everything after it is a decline. Open with the piece that raises the question and place the strongest at about two thirds.

What you now know

Liz Lerman's four steps put statements of meaning first, then let the maker ask the questions, then require responders to ask questions with no verdict hidden inside them, and only then release opinions, by permission. Alone, run describe, analyse, interpret and judge in that order, and refuse to judge first. Receiving well means not explaining, writing rather than arguing, splitting observations from prescriptions, and counting how many people independently noticed the same thing. Then select for what the investigation needs rather than for what you like, keep the failures that have consequences attached, and sequence so the work argues: by chronology, by variable, as question and answer, or by visual rhythm, but never best first.

That is the course. You began by choosing a question worth a year and learning to see edges rather than symbols, to measure rather than guess, and to build form out of value. You constructed space with vanishing points and with aerial perspective, and you learned to mix colour from six pigments rather than buy it in twenty four tubes. You composed deliberately, chose materials for the marks they make, and cut a block that printed the same image many times. You joined unlike images with scissors, took control of the camera in your pocket, and opened files in software that costs nothing. You found out what you may take from a source and what you owe it, how to photograph and describe what you made, and how to hear what other people see in it.

What remains is not on this page. It is the fifteen pieces themselves, made over months, each one changing something the last one revealed. Nobody can do that part for you, and no course can watch you do it. Find one other person to show work to regularly, keep the question in front of you, date everything, and keep the failures. Go and make the work.

Sources

1. Wikipedia contributors. (2024). Liz Lerman. Wikipedia. en.wikipedia.org
2. Wikipedia contributors. (2024). Art criticism. Wikipedia. en.wikipedia.org
3. Kollwitz, K. (1922-23). Widow II, from the portfolio War (Krieg) [Woodcut]. The Metropolitan Museum of Art, New York. metmuseum.org
4. Monet, C. (1894). Rouen Cathedral: The Portal (Sunlight) [Oil on canvas]. The Metropolitan Museum of Art, New York. metmuseum.org
5. College Board. AP Art and Design: course and exam description, sustained investigation and selected works. AP Central. apcentral.collegeboard.org