Questioning Direct Realism
First look at John Searle's case for direct realism
In his book, Seeing Things As They Are [Oxford Press, 2015] John Searle says that objects in the observer-independent world cause our subjective visual perceptions. And that as a consequence, seeing is knowing.
For example, when I look at a mountain, the mountain is a real object which causes my visual perception. The difference between this and hallucinating a mountain is that the hallucination is a subjective perception which was not caused by an actual external mountain out there.
Sounds good, right?
But what about when I see a mountain in a painting or a photograph? I’m sure Searle would agree that this is not an hallucination. And yet, it’s also not an actual mountain out there in the observer-independent world. So what in the photo caused me to perceive a mountain object?
The answer has to be that the same process is occurring when I see an external mountain as when I see a photograph (or holograph) of a mountain. My biological visual system creates a mountain object in each case from the photons sampled at my retinas.
In Searle’s account, the primary (and I guess secondary properties) of the physical mountain cause my perception of the mountain as an object. But in the case of the photo or hologram, there is no mountain with primary or secondary mountain properties. So the properties of the photo mountain or hologram mountain aren’t causing my mountain perception, because they aren’t mountain properties at all.
Instead, my brain creates the mountain object when sufficient patterns of photons are sampled by my retinas. As we know, a painting or photograph or hologram of a mountain is designed to create a sufficient pattern of photons to entice my brain to create a mountain object.
So I think Searle simply has it wrong.
Of course we know why he makes this error. Like all of us external-world realists, Searle wants to be able to say that there really is an existing external world out there.
Direct realism (seeing is knowing) allows us to confirm the common-sense view we all have that we are seeing an actual world when we open our eyes, and that objects in that world are directly causing our perceptions. We are designed by our evolutionary history to experience vision this way—it is preeminently useful to us as organisms that we see a world of objects and interact with them.
But direct realism can’t actually be the case.
Paintings, photographs, holograms, virtual reality headsets demonstrate the problem. They enable us to see objects that don’t actually exist as such. And in doing so, they reveal that we don’t actually see “things as they are”.
Objects are constructed by the brain. And so are their properties.
Two Questions
(1) So how do we know what, if anything, is really out there in the observer-independent world?
(2) How do we know that there is an observer-independent world?
The second question is trivially answered if naturalism is correct.
Since we are physical ourselves, and since our brains create our sensations and perceptions from sampling physical stuff (photons, we now understand) external to us, and since consciousness and all its accoutrements are produced by the brain and not located outside the brain, there must be an actual external world of non-consciousness stuff, which we call the physical world.
(Bishop Berkeley tried to present an alternative in which God’s consciousness is the source of an apparent non-consciousness world around us, and this possible alternative needs to be considered and evaluated, but the much simpler alternative is to suppose an external world of non-conscious physical stuff—it simplifies everything.)
That’s our answer to the second question.
The answer to the first question is pragmatic empiricism.
Imagine our sensations and perceptions to be a brain-produced stand-in for a physical world we can’t directly or even indirectly perceive. Immediately we face the question: how can we know anything about this external world?
First, the stand-in is our knowing of it. The stand-in is presented to us as a surrounding world.
We move and act within this apparent surrounding world, and since it turns out there actually is a real, unknowable, imperceptible external and surrounding world, the stand-in serves its purpose pretty well. Evolution has ensured this.
The brain pragmatically constructs the stand-in as a simulacrum which combines sensual experiences (vision, hearing, touch, etc) with construction of objects and their properties, and associated understandings (meanings/acquired knowledge) which make the stand-in increasingly useful to us as organisms.
What is the “criteria of truth”, that is, what makes the simulacrum fit for purpose, so that it more or less usefully stands in for the external world?
It’s this: that we are always pragmatically updating the simulacrum (updating our understandings and meanings) to improve its usefulness to us.
The scientific endeavor takes this a step further, adding instrumentation to extend our sensing (sampling) capabilities and applying a collective concept of usefulness to achieve point-of-view invariance1 in that regard.
I call this naturalism based on scepticism (which of course means philosophical/Humean scepticism), but you can see that it’s based on pragmatism as well. In my mind, pragmatic empiricism is the best description of how we go about fitting our simulacrum to the otherwise unknowable physical world.
Additional Evidence
Producing consciousness is a slower brain process. It doesn’t happen simultaneously with our sampling interactions with the world around us. It takes half a second or so for the brain to produce our vision and other conscious experiences and then react. Fortunately the brain can react faster than this, and in emergencies it will. For example, you might touch a hot stove and almost instantly retreat you hand before feeling a burning sensation or seeing and registering that your hand touched the stove. The brain does this non-consciously, which is faster.
A professional baseball batter can’t rely on the simulacrum to swing the bat because vision is too slow. The half-second reaction delay will not lead to success. But although consciousness is a slower brain process, it allows us to rewire our neurons (we call it learning), and the brain can now use those new neuronal connections to react quickly (outside of consciousness).
The process of learning to drive a car demonstrates the same thing. A beginning driver is hyper-conscious of everything (perhaps stressfully so) until their constant conscious effort has rewired the brain enough that aspects of driving can be handled outside of consciousness. Eventually driving becomes “second nature” in the sense that most of the initial hyper-consciousness is no longer necessary.
In short, learning a skill requires the effort of consciousness, but once a skill is learned the brain can forego most of that consciousness, and thereby speed up its actions and reactions. In this we start to see one of the key roles of consciousness: it is to rewire the neurons in the brain so that future actions can be done more quickly, outside of (by-passing) active “attention” consciousness.
Thus consciousness is employed by the brain to improve not just the usefulness of the stand-in (its fit to the external world), but to improve (and speed up) future actions and reactions within that world.
Back to John Searle
How can Searle address the issue of drawings, paintings, photographs, VR headsets causing us to experience the same “objects” we experience when directly looking at the external world?
He might do so by stressing “the distinction between object and content”2 Searle maintains that those who advocate alternatives to direct realism often fail to make this distinction.
For example, if I see a man in front of me, the content is that there is a man in front of me. The object is the man himself. If I am having a corresponding hallucination, the perceptual experience has a content, but no object. The content can be exactly the same in the two cases, but the presence of a content does not imply the presence of an object.3
Obviously, seeing the photograph of a man or seeing a VR image of a man cannot be considered (in Searle’s sense) an hallucination. So there will be a corresponding object for the perceptual experience of seeing a man. But here, Searle can say that the object is not a man, per se, but rather a photograph of a man or a hologram of a man, etc.
So far so good.
But notice now that there is a causation problem. Searle maintains, when we are not hallucinating, that objects in the external world cause our perceptual experiences.
But it seems to me that the existence of photographs and VR headsets demonstrates that it’s not objects which cause our perceptual experiences: it’s photons. And the photons are decidedly not the objects we experience.
The issue I’m raising here is not how do we distinguish between a virtual reality object and a real-world object, though that is an important issue. It’s rather, how can we seriously maintain that real-world objects cause our perceptual experiences, given this situation?
Yes, Searle is correct: as organisms we do believe by default that our perceptual experiences are caused by external objects in the world. But this is not in dispute.
What is in dispute is whether, in fact, external objects can actually cause the perceptual experiences we have of those “objects”. It seems obvious to me that they cannot, and that all perceptual experiences are in fact hallucinations.
At best our visual hallucinations are “caused” by the collection of photons at our retinas. Nevertheless, our visual hallucinations are produced by our brains, and populated with objects likewise produced by our brains.
Yet it’s photons, not objects out in the world, which trigger this chain of events.
Primary & Secondary Qualities
Searle’s brief discussion (in the last chapter) about Locke’s distinction between primary and secondary qualities is interesting. He notes, usefully, that
If you look at the lists of Primary and Secondary Qualities, you notice several things. First, each Secondary Quality is from one sense only. Colors are from sight, | sounds from hearing, odors from smell, tastes from taste. The Primary Qualities are accessible at two senses and always the same two, sight and touch. For each of the traditional five senses, there is one, and only one, Secondary Quality, with the exception of the sense of touch. That is why I think texture ought to be listed as a Secondary Quality, though Locke does not so list it. Why is it important to be accessible to two senses? And the answer is that it is part of our concept of a material object that it has these Primary Qualities and that our basic dealings with material objects are based on a coordination of sight and touch. When Locke says that they are inseparable from the object in whatsoever state it might be, he is getting at this conceptual point.4
My dictum is that whatever is “part of our concept of a material object” belongs to us and not to the external object. It’s constructed by our brain, not delivered to us by the object.
And this means that the “properties” of an object do not come from the object: they are created by our brains and experienced in consciousness, stored in memory.
Primary and Secondary Properties exist in here, not out there. Which means any “object” which is dependent upon its “properties” to exist (e.g. as a concept) exists in here, not out there. If we are to talk about objects out there, they will be objects without properties per se.
Consider this from a review of the Apple Vision Pro by Johnny Dodd for People magazine.
Then I spotted the virtual butterfly. It was flapping around in the distance, oblivious to the massive creatures nearby. At some point I stuck my hand out and watched it slowly land on my finger. And that’s when it happened. The moment its six spindly legs made contact with my finger, I could literally feel it touching my skin. The experience was gloriously perplexing and in the months that followed I convinced myself that I must have somehow imagined the whole thing.
Three weeks ago I was invited back to Apple’s L.A. headquarters to get some more time with the Vision Pro—and all I could think about was that damn butterfly. After putting the device through its paces and checking out some of the latest tweaks and new content, I asked my handlers if I could have one last encounter with their virtual winged insect. And sure enough, the moment it alighted upon my finger, it happened again. I could feel it.
And that’s when I realized that, of course, I was imagining the whole experience. That’s how powerful the Vision Pro is at re-creating the reality of meatspace (a.k.a the physical world). It serves up a jaw-droppingly realistic artificial world and my brain — doing what my brain always does — just filled in the gaps5
Our brains don’t perceive the physical world but rather create a sensual “experience” to stand in for the world, and this makes virtual reality devices possible. Ask yourself, how did Dodd’s brain “perceive” a butterfly landing on his finger from the digital 0’s and 1’s processed by the M2 chip in the Vision Pro?
It didn’t.
Rather, his brain constructed the butterfly on the finger (and the tactile sensation that went along with it) from his retina’s photon sampling. The M2 feeds the 0’s and 1’s to the OLED displays in the headset and this turns on and off millions of photon-producing pixels which then get sampled by the eyes.
No more nor less than what the brain is always doing as we move about in the world.
Species-Specific Vision
Photon-producing screens, whether CRT, LED, OLED etc, are designed and manufactured specifically for the benefit of human photon sampling. They are fine-tuned to meet our hominid-brain requirements. This makes it likely, indeed certain, that these photon-producing screens won’t work as well for other species.
If your pets seem disinterested in the tv in your living room, remember that it probably doesn’t produce the color range their brains expect when sampling photons. RGB works for the cones existing in human eyes, but other species have additional types of cones which collect photons our eyes can’t.
Even more importantly, their brains may use different rules to construct scenes.
Donald Hoffman identified 35 rules6 the human brain uses to construct vision. The less evolutionary history we have with another species, the more likely it is that their rules for constructing vision will differ from ours.
The Vision Pro is just the latest in a long line of human inventions designed to “trick” the human brain into seeing virtual (or imaginary) objects. The earliest cave paintings were an ancient equivalent to the Vision Pro. They worked, because the brain doesn’t just create a visual scene. It populates that scene with objects.
Draw an outline of a horse in the sand, and my brain doesn’t see a meaningless squiggle, it sees a horse. And it recalls horse-related meanings.
This tells us that the brain applies (or at least tries to apply) an “objectification” process to everything we see. We don’t see a meaningless scene. We see a scene of objects and potential objects. This is what makes vision useful to the organism: it creates meanings, and meanings suggest actions.7
See Victor Stenger, The Comprehensible Cosmos: Where Do the Laws of Physics Come From?, Prometheus Books, 2006.
John Searle, Seeing Things As They Are, Oxford Press, 2015, p. 35.
Ibid., p. 35.
Ibid., pp. 233-234
Donald Hoffman, Visual Intelligence, Norton & Company, New York, 1998.
Technically, meanings trigger feelings and the nature and intensity of those feelings suggest actions—although some automatic brain reactions bypass the feeling step to maximize the organism’s responsiveness (this is the goal of learning a new skill).
Somewhat related discussion in today's Jerry Coyne posting, q.v.
Dwight, not a Coyne column on gender—one on perceptions of mountains, etc. I’ll see if I can find it