From Cave to Cave: on Steven Pinker's "The Stuff of Thought"

NOTE: this is an old essay, and probably needs some serious reworking.  It's really just here in case you're bored and desperate, in which case, well, it can't make things any worse.  Probably.

For anyone interested in language, The Stuff of Thought is a great pleasure to read.  Watching Steven Pinker delve into language can be like watching a magician work – he pulls fascinating insights from your language as unexpectedly as a magician pulls the five of hearts from your pocket.  He shows you layers of meaning and fine-tuned distinctions that solve conundrums you never even thought of.  These linguistic investigations, though, are only a means to an end in this book: Pinker’s ambitious goal is to use language to show how we think about the world.  His major claim is that there are conceptions of reality inherent in human language, and that “they add up to a distinctively human model of reality, which differs in major ways from the objective understanding of reality eked out by our best science and logic.”(vii) 

How far Pinker succeeds in doing this is debatable.  There are moments when, after a chapter full of dazzling linguistic insights, it is difficult to pin down what those insights fundamentally say about how we think.  And Pinker’s explanations of these implications are not always helpful.  The reason for this – and the fundamental problem with the book – is that there are underlying flaws in the premises of Pinker’s investigation.  In the quotation above, Pinker makes a distinction between two models of reality.  This distinction is a sine qua non for the argument of the book.  Were the concepts of reality that Pinker extracts from language to correspond precisely with objective reality, they could tell us nothing about human nature – they would simply tell us that language is a clear lens, and looking at reality through it is nothing but looking at reality.  For the premise of the book to be valid, there must be a difference.  The concepts of reality in language must show us reality through the lens of the human mind, and by comparing that reality to objective reality, we must be able to come to certain conclusions about the lens.  The trouble is that Pinker not only assumes that there is objective reality, but that it is precisely the reality depicted by human science and logic.  In actuality, science is also a lens through which we view reality, and it is a product of the same human minds as language.  This fundamentally undermines Pinker’s method of investigation in the book. 

In the body of the text, this problem manifests itself as a troubling resemblance (which Pinker avoids mentioning) between the models of reality implied by language and the corresponding “objective” scientific models.  We can see this most clearly in chapter four (“Cleaving the Air”), where Pinker examines the models of time, space, substance, and causality that underlie our speech.  Let’s start with examining time. Pinker notes that our language treats time in two peculiar ways: it lumps it into chunks of events and segments rather than viewing it as one continuous stream, and it treats time as space.  So we discuss as discreet entities the past, present, and future, and that time last year when Jenna stepped on a porcupine; and we say that Jack moved the meeting time back a few hours, or that the deadline is not as far away as we thought.  But are these conceptions of time really so different from a scientific model of the world?  In the first case, it’s important to note that our conception of time is not really segmented.  We refer to the past to differentiate it from the part of time that we haven’t yet experienced, but it’s not that we’re actually segmenting time – if we were, we wouldn’t use the spatial metaphors of back and far ahead, which construe time as a continuous line.  And the fact that we sometimes refer to one block of time as its own entity isn’t strange or unscientific – this is a perfectly scientific practice when focusing on a specific motion, such as a rock dropping, that has a definite beginning and end point.  The fact that a scientist does not deal with the time before or after the rock’s fall doesn’t mean that the scientific model has a faulty view of time; rather, like the linguistic model, it exercises selective focus.  And viewing time as space, which Pinker calls a metaphor (6), may not be a metaphor at all.  Since Einstein’s theory of relativity and subsequent developments in physics, physicists have viewed time as a dimension, just like the three dimensions of space.  So the fact that language treats all four of these dimensions similarly does not conflict with scientific reality – in fact, it was the linguistic model of reality that had it right all along, while the scientific model didn’t catch up until the last century.

As for space itself, Pinker notes that our conception of space as illustrated in language is painfully inexact and extremely oversimplified.  Our prepositions do not measure distance at all accurately, and they often idealize real three-dimensional shapes into zero-, one-, or two- dimensional shapes.  For example, prepositions such as in, on, or near treat their object as a zero-dimensional point, and across treats its object as a one-dimensional line.  The inexactitude of distance is a valid difference between the model of language and the model of science.  However, the simplification is not.  When astronomers calculate the distance between earth and various stars, they treat both earth and the stars as points.  In one sense, this is ludicrous: the earth is so large that it contains all human life, and each star is larger by several orders of magnitude, and yet the scientists are brushing both off as points.  But in this context where the astronomers are focusing on colossal distances, they simplify the shapes that are not important.  The tendency of language to focus on one aspect of space and simplify others is not unparalleled in the scientific model of spatial relations.

We see the same principle of focus and simplification at work in both language’s and science’s conception of substance.  Pinker points out how we often construe a group of individuals – such as pebbles – as an amorphous stuff, like gravel.  This certainly seems arbitrary and imprecise at first.  But every physical science that doesn’t operate on a subatomic level is using the exact same principle.  In biology we generally study living things made of matter – a fancy way of saying stuff – and not large numbers of subatomic particles.  The study of fluid dynamics is the study of how moving, liquid stuff behaves.  Aerodynamics studies how differently shaped objects (made of – you guessed it – stuff) behave in the stuff we call air.  So either the scientific model of reality is subject to the same imprecision, or language’s view of groups of individuals as stuff isn’t so peculiar after all.

The final principle inherent in our language that Pinker discusses in this chapter is causality.  Verbs such as pour and spray use different constructions because pour means letting gravity act on the water, whereas spray means actively causing the water to spray.  Pinker notes that causality is a troubling concept and leads to all sorts of paradoxes, especially in legal and moral cases where ultimate responsibility must be determined.  He even mentions how some scientists and logicians (most notably Bertrand Russell) have declared that causality does not exist.  He shows that our intuitive view of physics as expressed in language involves episodes of forcing and helping and prohibiting, whereas in actuality physics is just a bunch of equations.  But this is nonsense.  The equations used by physicists express how the gravitation of the sun influences the earth to move in an elliptical path; even if the physicists do not explicitly call this causation, it is precisely what we mean by causation in an everyday sense.  So the principle of causation may be oversimplified in language, but one cannot write the whole idea off as a quirk of our intuitions, since it is mirrored in science.

Now let’s step back from these specifics for a moment and look at Pinker’s general conclusion about the shortcomings of language in describing objective scientific reality.  He says that models of reality in language “are digital where the world is analogue, austere and schematic where the world is rich and textured, vague even when we crave precision, and parochial to human goals and interests even when we ought to seek the view from nowhere.”(233)  The striking thing about this description is that, aside from the phrase “vague when we crave precision,” every aspect of it perfectly describes the scientific view of reality.  Science is digital where the world is analogue, for science measures everything, and all measurement is approximation.  Science is austere and schematic with its definitions and equations and models, whereas the world is rich and textured.  Science is by definition parochial to human goals and interests, since some human was interested enough to study everything that has been studied.  So while Pinker is trying to critique language’s depiction of scientific reality, he is equally using language to critique science’s depiction of reality.  

This brings us back to the distinction between objective reality and reality as we see it through science.  Just like the reality we see through our intuitive use of language, scientific reality is reality through a lens.  Pinker says that “Our intuitions of these entities are riddled with paradoxes and inconsistencies.  But reality [italics Pinker’s] can’t be riddled with paradoxes and inconsistencies; reality just is.”(155)  He is right, of course, that reality cannot be self-contradictory.  But of course the trouble with this is that the scientific view of the reality is constantly subject to new paradoxes and inconstancies, which must be resolved by new theories.  So again we see the gap between the scientific view of the reality and reality itself, a gap which – though it may be smaller than the gap between our intuitive model of reality and reality itself – is still considerable.  The scientific lens is much more polished, and it refines our intuitions about the world with many levels of reasoning and observation, but it is still a lens and shares the same source of humans’ reasonable inferences about reality.

So where does this leave us?  Pinker concludes the book with a reference to Plato’s allegory of the cave.  He construes our intuitive model of reality, which he has shown through the examination of our language, as the shadow-play on the wall of the cave.  He suggests that we can begin to feel our way out of the cave – transcend the limitations of our intuitions – and see true scientific reality by using metaphor and the infinite combinatorial possibilities of language.  But Pinker has not demonstrated how our intuitive model of reality differs from objective reality, only how it differs from our scientific view of reality.  And the scientific model of reality is not the open sunlit earth; it is another cave.  The shadow-play may be a bit sharper, but it is not reality any more than the shadow-play in the first cave is.  Now, this is not to say that it is useless to make what distinctions can accurately be made between the two models, or that nothing can be gained by looking to the scientific model.  Two caves, like two heads, are better than one.  But we must be a bit humbler than Pinker would have us be: even in the second cave, we are still cavemen.  If the view from nowhere does exist, science is certainly not it; and though metaphor and combinatorial language can help us transcend our intuitions and feel our way into the second cave and perhaps beyond, we are still far from daylight and still questioning what sort of thing daylight could be.