Subject: Who the Photograph Is About

Show someone a photograph of yours and ask: “what did you see first?”

If their answer is not what you had in mind, that is not their failing.

This chapter is about closing that gap, and the first step is accepting something uncomfortable: where the viewer looks is not up to you.

The question to ask on location is this:

If only one thing could stay in the frame, which one?

Until you can answer it, it is not yet time to press the shutter.

The question is not the same as composition. Composition asks where things go; subject asks why it is this thing at all. Part two taught the first, and the first only means anything once the second has an answer — a subject you are unsure of stays unsure when you place it on a third.

Your eye is not looking at the whole photograph

The mechanism first, because it explains every rule that follows.

Only a small central patch of the retina — the fovea — resolves fine detail, covering roughly two degrees of vision, about a thumbnail at arm’s length. Resolution collapses rapidly outside it.

So “I am looking at this photograph” is an illusion. What actually happens is that the eye jumps three or four times a second, resting for two to three hundred milliseconds at a time, and the impression of a whole picture is assembled from a sequence of small patches. The jumps are called saccades, and vision is suppressed during them — you never see the movement itself.

Which makes the real question: how does the eye choose where to jump next?

It can only use the coarse information peripheral vision supplies, and peripheral vision has definite characteristics:

• Highly sensitive to brightness differences and to motion — evolutionarily the two things that mattered.
• Poor at detail.
• Poor at colour too — cones are concentrated in the fovea and the periphery is nearly all rods.

That already yields a strong practical conclusion: luminance decides where the viewer looks; hue does not. The previous part’s rule about a small saturated accent carried an unstated condition — the accent must also differ in brightness, or be large enough for peripheral vision to register. A saturated patch at the same luminance as its surroundings is close to invisible out there.

Visual weight, though, only settles the first look. In the 1960s the Soviet researcher Alfred Yarbus ran a now-classic experiment: the same viewer, the same painting, but a different question each time — how wealthy is this family, how old are the people, memorise where everyone is standing. The recorded scan paths barely resemble one another.

Both halves of that matter. The first two or three jumps are set by the physical properties of the picture and are yours to control; everything after is set by the question in the viewer’s head and is not. What composition can actually do is make sure the first landing is the right one — because that landing determines which question the viewer carries into the rest of the frame.

This is worth more than it sounds, because most photographs today are not contemplated. They are met while scrolling, in the second or two before a thumb moves — a handful of saccades in total. For most of your audience, the first fixation is not the beginning of the encounter. It is very nearly the whole of it.

Faces are the one exception, and it is an exception at the hardware level. A dedicated region of the temporal lobe processes them, fast enough to finish before you are aware of it, and crude arrangements of two dots and a line will set it off — which is why you see faces in power sockets, car fronts and clouds.

The consequence is blunt: any face in the frame is the subject, whether you wanted it or not. The stranger walking past in the background counts.

Visual weight

The mechanism, arranged as something usable:

What adds weight	Why it works	How you use it	How it hurts you
Faces, eyes	Dedicated neural hardware, too fast to resist	Make the subject the only face	Passers-by, faces on posters
The brightest region	Peripheral vision’s strongest cue	Put the subject in the light	A blown corner of sky, a window
The highest-contrast edge	Saccade targets are chosen by edge density	Separate subject from background in luminance	Railings, branches, wire fencing
The sharpest region	Sharp things get foveal confirmation first	Focus only on the subject	Something behind the subject is sharper
The most saturated colour	Strong in the fovea, weak in the periphery	Make the subject the only saturated thing	Signage, bins, advertising
Whatever differs most	Difference is itself a signal	One red note in grey, one diagonal among verticals	The accidental exception — a crooked sign
Gaze and implied motion	We automatically follow what others look at	Have people in frame look at the subject	The subject looks out of frame and takes the viewer with them

The “no clear subject” complaint is almost always one row of this table landing in the wrong place. And note the right-hand column: every tool here is simultaneously a trap.

Letting the subject win

There are only two families of method: add weight to it, or take weight off everything else.

To add weight to the subject	To take it off everything else
Put it in the brightest light	Let the background fall into shade
Separate it from the background in luminance	Open the aperture and blur the background
Focus on it alone	Move closer and exclude the clutter
Lead the eye to it with lines	Wait for the clutter to leave
Give it space around it	Find a clean wall to put behind it

The two columns are not equally effective. Subtraction almost always wins, because visual weight is relative — darkening the background by a stop is equivalent to brightening the subject by one, without touching the subject’s texture or skin.

And within subtraction, moving closer is the most underrated move there is. Beginners photograph from the middle distance, where the frame contains a little of everything and enough of nothing. Three steps forward solves several problems at once: clutter leaves the frame, the subject’s relative area grows, and the fall-off of the light becomes visible. One action, three fixes, no equipment.

The subject need not be large

A small subject surrounded by emptiness can carry enormous weight, because isolation is a form of contrast. Peripheral vision sees a large unvarying field, and the single variation in the middle of it becomes the only place worth jumping to.

The test is not what fraction of the frame the subject occupies. It is whether the relationship between subject and surroundings is what the photograph is about. If the surroundings are helping to say something, keep them; if they merely happen to be there, cut them.

This is also where empty space is most often misused. Space is not “nothing beside the subject”. For it to work, the emptiness has to mean something — exposure, isolation, waiting, insignificance. Emptiness without a meaning is just frame you paid for and did not use.

The subject need not be a thing

Everything so far has assumed the subject is an object — a person, a building, a bird. Many good photographs have subjects that are not objects:

• A relationship. The distance between two people, the ratio of a person to a building. No single element is the protagonist here; the protagonist is the space between them, and that space has to be kept clean, with nothing else allowed into it.
• A quality of light. The whole of the part on light. Who the light falls on is secondary; what matters is the shape, edge and direction of the light itself.
• A condition. Rain just stopped, the crowd just left, the fog has not lifted. This kind of subject has no boundary — it is distributed across the whole frame.

This has a practical consequence: when the subject is a relationship or a condition, “move closer” stops working and can actively damage the picture. Moving in severs the relationship and crops away the surroundings the condition depends on. Which returns to the same sentence: decide what you are saying first, and only then can a rule help you.

The focus point is not the subject

A common confusion worth separating: focusing is a technical act, choosing a subject is a narrative one. They usually coincide. They are not the same level of decision.

Shallow depth of field does add visual weight — sharp things get foveal confirmation first, as the table said. But it has two side effects.

First, it discards the information in the surroundings along with the clutter. If the photograph is about a person’s relationship to a place, blurring the place past recognition deletes half the subject. Wide apertures often make pictures look more professional because they force a subtraction; that subtraction is not necessarily the one you wanted.

Second, it lets you off the hook. Blurring a messy background at f/1.4 is much easier than walking twenty steps to find a clean wall, and the results differ: blurred clutter is still clutter, now in the form of coloured, bright blobs that go on holding weight in peripheral vision. A genuinely clean background contains nothing, not something you cannot make out.

A suggested order of operations: solve the background with your feet, and give the aperture only what is left.

One photograph, one thing

The last principle, and the hardest: a photograph says one thing.

Trying to convey expression, atmosphere, scale and weather at once usually leaves all four vague. Better to make four photographs — which is exactly what the next two chapters are about.

How do you know whether what you want to say is one thing? A serviceable test: say it in one sentence that contains no “and”. “A mother is worried” is one thing. “A mother is worried, and the camp is squalid, and this is the Depression” is three. They are all true, and one photograph will not hold three.

Six common symptoms and their causes

Always: the picture shows X → the usual cause is Y → change Z.

1. The first thing people see is not your subject → one row of the table has landed in the wrong place, usually “the brightest region” → find the brightest patch in the frame; if it is not the subject, darken it or exclude it.
2. A stranger in the background has taken the photograph over → faces are a hardware-level priority and do not distinguish lead from extra → there is no compromise: wait for them to leave, change angle, or blur them past being recognisable as a face.
3. The picture looks rich but says nothing in particular → too many elements at similar weight, none clearly winning → subtract: move closer, use a longer lens, wait for some of it to leave. Adding will not fix this.
4. The subject is large and still does not stand out → size is not the main source of visual weight; luminance and contrast are → separate subject and background in brightness, or find a cleaner background.
5. Plenty of empty space and it reads as merely empty → the emptiness has no meaning and is not part of the content → ask what it is saying; if you cannot answer, crop it and use the close version instead.
6. The subject looks out of frame and the viewer drifts out too → implied gaze is a powerful lead and it is currently pointing outside → leave space on the side they are looking towards, or wait for them to turn back. Cropping tight on the gaze side makes the frame tense — a deliberate effect, not a default.

Lange made six frames that day

The hero of this chapter is Dorothea Lange’s Migrant Mother, 1936. It is close to a textbook case of all the visual weight stacking in one place, but what is worth learning is how she arrived at that frame.

She had stopped at a pea-pickers’ camp at Nipomo, California. By her own later account she made five exposures — the Library of Congress holds six, and further frames survive in other collections — and whether five or six, they are a very short and visibly converging sequence: from a wider frame including the tent and its surroundings, progressively closer, until the last one is the composition we know — the mother and three children, with the background almost entirely excluded.

That process is this chapter’s two rules being executed:

• Move closer. Every step removed information about the setting and increased the mother’s relative weight.
• One photograph, one thing. The wider frames are simultaneously about the camp, the condition of migrant labour, and this family. None of them is bad. None of them is as certain about who it is about.

One further detail is worth knowing. Those photographs were made for the federal Resettlement Administration — reorganised in 1937 into the Farm Security Administration everyone now names — whose photographic section was directed by Roy Stryker, who issued his photographers shooting scripts — not “go and make some pictures” but explicit lists of what was to be documented. In that programme, “what are you actually saying” was written down before anyone picked up a camera. The order of this chapter — decide who first, arrange the frame second — is not a modern teaching device. It is how those photographs were made.

(And because they were the output of a government programme, they are in the public domain today, which is why this site can use them — and why the same names keep appearing throughout this part.)

A table for the field

What to confirm	How to check on the spot	What to do if it fails
Is the subject the brightest thing?	Squint until detail goes; see which patch stays brightest	Darken the rest, or wait for the light to move
How many faces are in frame?	Count them, including background and posters	More than one: get it down to one
Is the subject clipped?	Check all four edges	Step back half a pace, or turn the camera
Does the empty space mean something?	Say aloud what it is doing	If you cannot, move closer
Where does the gaze go?	Follow the direction people in frame are looking	Pointing out of frame: leave space on that side
What is this about?	Say it in one sentence	If it takes two, make it two photographs