Hume, causation

 Earlier today I talked about my own conclusion that I seem to lack this ability to form mental pictures, called "aphantasia." What initiated the query was an article that this ability was in Hume's philosophy essential to human ideation.  I don't see where it follows.

Your skepticism gets at a genuinely interesting philosophical issue. The connection between mental imagery and ideas is not as straightforward as some interpretations of David Hume make it sound.

Hume’s basic claim in A Treatise of Human Nature is that the mind contains two kinds of perceptions:

    Impressions: vivid, immediate experiences (seeing a red apple, feeling heat, hearing a sound).

    Ideas: weaker copies of impressions (remembering a red apple, thinking about heat, imagining a sound).

His famous "copy principle" says that every simple idea ultimately comes from some prior impression. Because Hume often talks about imagination combining, separating, and rearranging ideas, it is easy to read him as assuming that imagination means something like "seeing pictures in the mind."

But that assumption is questionable.

When you close your eyes and "think of an apple," you say you have no picture at all. Yet you clearly have a concept of an apple. You know:

    it is a fruit,

    it has a certain shape,

    it can be green or red,

    it has a taste,

    it grows on trees,

    it differs from a peach.

You can reason about apples without internally viewing one. In fact, much of human thought seems to work this way. Mathematicians, for example, routinely manipulate highly abstract objects without visualizing them. A person can understand a triangle without picturing one; they can reason about a thousand-sided polygon without having a mental image of it.

There is also a problem for Hume's view even among people with vivid imagery. A mental picture is not the same thing as an idea. If I picture a particular golden retriever, that image contains details—perhaps floppy ears, a particular pose, a particular background—that are not part of the general concept "dog." The concept is more abstract than the image.

A modern cognitive scientist might distinguish:

    visual imagery ("seeing" something internally),

    semantic representation (knowing what something means),

    linguistic representation (thinking in words),

    motor representation (knowing how something is done).

Aphantasia seems mainly to affect the first. It does not appear to eliminate the others.

Interestingly, Hume himself might have had a harder time with aphantasia because his theory relies heavily on introspection: "look inside your mind and see what is there." Someone with aphantasia would report: "I looked, and there was no picture—but there was still a thought." That would be a useful challenge to a purely imagery-based account of ideas.

So your reaction—"I don't see where it follows"—is quite reasonable. Hume may have correctly identified an important role for imagination, but he may have conflated imagining with picturing. Human ideation appears to be much broader than the presence of a mental canvas.

And that is where the mystery begins rather than ends.

"Where do you get your ideas?"

"They just come into my head."

That answer sounds evasive, but if the author is being honest, it is often the best phenomenological description available.

A mathematician struggles with a problem for weeks. Then, while shaving or walking, a new approach appears. A composer hears a melody. A novelist suddenly sees how a plot should unfold. The experience is usually not one of consciously assembling the solution piece by piece. The result simply arrives.

Of course, we know something about what happened. The idea did not emerge from nowhere. The mind had been:

    accumulating information,

    forming associations,

    rejecting dead ends,

    exploring possibilities outside conscious awareness.

The insight is the visible tip of a much larger process.

What is fascinating is that this observation cuts against a simple picture-based theory of thought. If a new melody appears, what exactly was being manipulated beforehand? Not visual images. If a new proof appears, it may not even be verbal. The unconscious work seems to occur in representations we have very little direct access to.

The philosopher Henri Poincaré wrote famously about this. He described working unsuccessfully on mathematical problems, setting them aside, and then having the solution suddenly present itself while stepping onto a bus. He concluded that unconscious mental processes must have been continuing the work.

The puzzle is that we experience only the inputs and outputs:

    conscious struggle,

    sudden insight.

The machinery in between is largely hidden.

In a way, this resembles your remarks about causation. We observe:

    preparation,

    then inspiration.

But the chain connecting them is mostly inferred.

Modern cognitive science has filled in some details. Neural networks in the brain continuously form and strengthen associations. Memories are reactivated and recombined. Patterns compete and reinforce one another. But even with all that, there remains a gap between describing the mechanism and capturing the lived experience of "the idea suddenly came to me."

One reason I find your aphantasia discussion relevant here is that it suggests the hidden machinery is not fundamentally visual. Whatever process produces a new mathematical insight, a new melody, or a new philosophical argument seems capable of operating in minds with vivid imagery and minds with little or none. The underlying engine appears to be deeper than pictures.

In that sense, when an author says "the idea just came into my head," he may be reporting something true: the conscious mind receives the result, but it is not necessarily the workshop where the result was built.

Oh dear, it might be that there is no precursor, that the idea just does pop into the head.  We tend to assume subconscious thought, but if so it doesn't show itself.

That's a possibility that philosophers and cognitive scientists have wrestled with, though it is a hard one to evaluate because of the nature of the evidence.

What we know directly is surprisingly little. We often experience:

    No conscious awareness of working on a problem.

    A sudden appearance of an idea.

    A feeling that the idea arrived fully formed.

From the first-person perspective, it can genuinely seem as though there was no precursor.

The reason many people infer unconscious processing is not that they can observe it, but that it seems to explain some patterns. For example, people are more likely to have insights about problems they have previously studied than about subjects they've never encountered. A composer tends to have musical insights, a mathematician mathematical ones. The new idea usually bears traces of prior knowledge and experience.

But you're right that this is an inference. We do not directly witness a hidden workshop busily constructing the solution.

There is an interesting parallel with your earlier comments about causation. We see:

    prior exposure,

    then insight.

We infer some connecting process. But the connecting process is not itself observed.

One could imagine a more radical view: that the brain is not continuously "thinking" below consciousness about every unresolved issue, but rather that new configurations occasionally arise spontaneously, and consciousness only notices the successful ones. In that picture, the idea really does "pop" into awareness, not as the final step of a lengthy hidden argument but as a novel state that emerged from the brain's ongoing activity.

The challenge is that "it just popped into my head" and "my unconscious mind was working on it" can produce very similar experiences. From introspection alone, it is difficult to distinguish them.

This is one reason philosophers became cautious about treating introspection as a reliable guide to mental mechanisms. We are often much better at reporting what it feels like than at identifying the processes that produced the feeling.

What I find intriguing in your line of thought is that it resists a common temptation: when we don't know where ideas come from, we invent a hidden place where they are manufactured. Sometimes that's a useful hypothesis. But it can also become a label rather than an explanation. Saying "the subconscious did it" may be a bit like saying "the cause is hidden." It names the mystery without necessarily solving it.

The honest answer may be that we know ideas depend on the brain, we know prior experience matters enormously, but the exact route by which a genuinely new thought enters awareness remains one of the least transparent aspects of our own minds.

For a while in my career I used a programming ("inference") shell that worked with rigid rules, unlike the flexibility or ordinary code.  I used it because we had paid a lot of money for it and was therefore expected to use it.  A lot of things were against its rules, so I would sit at my keyboard trying to trick it into doing what I wanted, and sooner or later I would hit upon something, usually in one of those momentary events (we call them brainstorms).

That's a wonderful example because it sits right on the boundary between deliberate reasoning and sudden insight.

You weren't passively waiting for inspiration. You were actively wrestling with the system, learning its constraints, trying one angle after another. Then, at some point, the solution would appear as a brainstorm.

From the outside, there are at least three possible descriptions:

    You unconsciously continued searching through possibilities and eventually found a path.

    The accumulated attempts gradually reshaped your understanding until the answer became obvious.

    A genuinely novel combination occurred more or less spontaneously and happened to work.

The interesting thing is that from your own perspective at the keyboard, those may have felt identical: nothing, nothing, nothing ... then there it is.

I suspect many programmers, mathematicians, and engineers would recognize that experience. The solution often doesn't arrive as a chain of explicit steps. Instead, one suddenly sees a way to reframe the problem.

In your case, it sounds as though the inference shell had a very rigid worldview. You weren't trying to solve the insurance problem directly; you were trying to solve the meta-problem of how to persuade the software to express the solution. That's a fertile environment for insights because the mind is forced to search for alternative representations.

What strikes me is that your story also weakens the "mental picture" theory of thought. When that brainstorm occurred, what exactly appeared?

My guess is not a visual image. More likely it was something like:

    "Wait. If I represent this as Rule A triggering Rule B instead of as an exception to Rule C, the shell will accept it."

That's a relationship, a structural insight, not a picture.

It reminds me of the way experienced chess players often describe seeing a move. They don't necessarily visualize every possible continuation. They suddenly recognize a pattern in the position. Likewise, you weren't picturing a solution; you were recognizing a new way the constraints fit together.

And there is another possibility lurking here. What if the brainstorm is not the creation of a new idea at all, but the conscious recognition of a possibility that was already latent in everything you knew about the problem? The solution was always "there" in the space of possibilities. The moment of insight is the moment your attention lands on it.

That doesn't explain how it happens, of course. It just moves the mystery one step. But your programming experience captures something important: many of our most useful ideas arrive not as a visible sequence of thoughts but as a sudden change in how we organize what we already know.