Freedom and Bondage in Psychology

Freedom and Bondage in Psychology

Chomsky has long urged that the use of language is stimulus-free. The point in itself is obvious, as many of the most important points are, but it stands opposed to entrenched ideas. I will make some remarks about its interpretation and significance. The notion of a stimulus was introduced in the nineteenth century in connection with psychophysics. Just as a physical cause leads to physical effects that can be studied and measured, so a physical cause leads to psychological effects that can be studied and measured. Just as we can establish quantitative physical laws relating physical magnitudes, so we can establish psychophysical laws that relate physical and psychological magnitudes. For example, we can investigate how visual sensations depend on physical properties of external objects, discovering that the intensity of the former varies with the intensity of the latter (The Weber-Fechner law). Here the stimulus is light and the response is a visual sensation. We thus arrive at the idea that physical stimuli produce and are lawfully correlated with sensations conceived as internal psychological occurrences: we can predict the latter from the former according to quantitative laws. The stimulus elicits the response—makes it happen in a predictable manner. This makes the relationship similar to that between purely physical causes and effects—as in Newton’s laws of motion. Just as mass and force lead to acceleration, so light intensity leads to visual vividness. We can thus model psychophysics on physics: both involve law-like relations between magnitudes, with stimulus and response acting like physical cause and effect. This gives the experimental psychologist something tangible to work on, namely mapping the laws that connect physical stimuli with psychological responses. A science can thereby be constructed.

            This basic model was extended by subsequent behavioral psychology: now the response is taken to be overt behavior not subjective quality, and the laws relate physical stimuli to observable behavior. But the basic theoretical model is the same: stimuli eliciting responses in a lawful quantitative manner. We might say that psychology, so conceived, is the science of the laws of elicitation—how stimuli elicit responses, reliably and measurably, just as physical causes elicit physical effects. Thus stimulus-response (S-R) psychology was born: the experimenter could vary the stimuli at will—their intensity, frequency, probability, and timing—and discover the properties of the response. S-R psychology is a kind of physics of the behaving subject. From this perspective it is not that the response is behavioral that is crucial (it wasn’t for psychophysics); it is that the response is something elicited by the stimulus—caused by it, predictable from it, tied to it. Hence theorists took to speaking of evokingresponses by stimuli or stimuli triggering responses: responses are stimulus-dependent, stimulus-bound, stimulus-controlled. There is no response but that a stimulus makes it so—nothing is free from stimulus bondage. It might take some ingenuity to find the stimulus, and we may have to accept merely stochastic dependencies, but fundamentally all behavior is determined by outside stimuli. The brain (mind) is accordingly an S-R mechanism—a device for transmitting stimuli to responses; it plays no autonomous role in generating behavior—everything traces to those eliciting stimuli. Everything psychological works in the same way as it does in psychophysics—this is the model and paradigm.

            It is against this background that Chomsky’s point is to be understood. For his point is that language use—speech acts, utterances—do not conform to this model, contrary to the precepts of S-R psychology. What a person says on an occasion is not controlled by the stimuli impinging on that person—not elicited, evoked, or triggered. It may be semantically related to the environment of the speaker, but it isn’t fixed by that environment—the person might say something completely different in an identical environment. Still less is the utterance shaped by the surrounding stimuli: it doesn’t vary according to variations in the stimulus, say by growing louder or softer, or more or less interesting. The properties of the “response” have nothing to do with the properties of the “stimulus”. We should drop talk of stimulus and response here altogether: the utterance is not a response to anything in the array of impinging energies. The S-R model is quite inappropriate to the case. It follows that the mind-brain is not an S-R mechanism, at least where language is concerned. It works according to quite different principles. Nor is this a matter of merely statistical S-R relations: there is no law at all connecting environment to utterance. The environment might suggest a remark to the speaker, but there is no sense in which it elicits a remark. By contrast, a physical stimulus can elicit a perceptual response, but it doesn’t “suggest” such a response. On no account should we conflate these two relations. Linguistic behavior therefore does not fit the S-R model, but requires a quite different theoretical treatment.

            All this strikes me as unexceptionable and I won’t defend it further. I am interested in extending the point beyond language and in exploring the resulting picture of the human mind (and maybe the minds of other animals). First, we should note that the point is not that all behavior or psychological response is stimulus-free: much of it is stimulus-bound and fits the S-R model. The patellar (knee) reflex is an obvious example, along with the blink reflex and the salivation reflex. Chomsky’s point is that speaking is not like that. Conditioned reflexes are the same: stimulus-bound not stimulus-free. Likewise for perceptual responses: the perception is elicited by the stimulus (distal or proximal) and its properties depend on those of the stimulus (though not exclusively). The perceptual systems are S-R systems (which is not to say they are simple or mechanical). What you see is what you are made to see not what you make to see; but you make utterances and are not made to. The mind is therefore composed of two sorts of system: bound and free. It is not all bound or all free, but a combination. Just as speaking is not like seeing, so seeing is not like speaking: seeing an object is not like speaking in its presence—it is not a kind of commentary on the passing show. It doesn’t have that kind of freedom; it is not a kind of language (“That looks a duck”). We must not make an error of assimilation in either direction. The mind is both reflexive and reflective; both bound and unbound.

            What else is stimulus-free? I suggest that thought is. What a person thinks is not elicited by his or her environment, though it may be suggested by it. Thoughts are not responses to environmental stimuli. A person can think indefinitely many different thoughts in the same stimulus situation, since thoughts do not occur because of triggering by the stimulus situation. Their causation (if that word is appropriate) is endogenous. Your thoughts can roam far and wide while your senses are stimulated thus and so. There is no supervenience of thought on outer stimuli. A stimulus may prompt a particular thought, but it may not; thought operates autonomously, without reliance on triggering stimuli. Beliefs may be elicited by perceptual stimuli, but thinking is another matter. So thought is like speech: stimulus-free and endogenously generated. Speech is an external action while thought is something internal, but both differ from perception and reflexes. In fact, I am inclined to suggest that language is stimulus-free because thought is stimulus-free: our thoughts lead to our speech acts (though not as stimulus to response) and our thoughts are stimulus-free, so our speech acts are likewise stimulus free. We say what we do because of what we think, but what we think is not controlled by impinging stimuli (unlike what we perceive); therefore free speech (in this sense) is underwritten by free thought. At some point in evolution the mind detached itself from S-R relations (a happy mutation!) and developed operational autonomy, thus producing thought; language then built upon that foundation, developing its own autonomy. The decoupling of world and mind is what stimulus freedom amounts to, and both thought and language share it. If we conceive of thought in terms of the language of thought, then the autonomy of thought rests upon the autonomy of the internal language—utterances in the language of thought are stimulus-free. But then it is the stimulus freedom of the inner mental utterance that is basic, not the outer vocal performance. In any case, language and thought are clearly intertwined when it comes to the impotence of the stimulus. And this is not a matter of the “poverty of the stimulus”, because thought and speech are not responses to stimuli at all, however poor and meager. They don’t result from an inadequate stimulus but from no stimulus. S-R psychology simply doesn’t apply to them.

            A certain kind of dualism takes shape around these observations—a dualism within the mind. Human minds (and maybe others) have two very different kinds of faculty, according as the faculty is stimulus-bound or stimulus-free. S-R psychology applies to one sort of faculty, including perception and motor reflexes (conditioned and unconditioned), but it fails to apply to another kind.  [1] We must not try to assimilate one kind to the other in either direction. We would therefore expect that different sorts of theory are appropriate to the two sorts of faculty, with one sort mirroring theories in physics and the other not. As I remarked, psychophysics modeled itself on Newtonian mechanics; but the theory of thought and speech will not be like that, because of inherent stimulus freedom. This prompts the question of why such a division exists: what is it about speech and thought that makes them stimulus-free? Is there something about their intrinsic structure that leads to their stimulus freedom? The obvious place to look is productivity—the potential infinity of sentences and thoughts grounded in finitely many combinable elements. Is this essential creativity the reason for stimulus freedom? Or are these two independent features of speech and thought? The natural hypothesis is that they could not be productive in the way they are if they were stimulus-bound. Suppose we try to imagine a speaking creature whose speech acts are as rigidly governed by eliciting stimuli as our reflexes and perceptions are. This creature cannot speak unless it is triggered into doing so by an impinging stimulus, as we cannot see without being stimulated to see. Its verbal behavior is subject to S-R psychology. And this is not just a motor limitation; it can’t even enunciate sentences in its head without an external stimulus. This is hard to get one’s mind around, so alien is it to our own relation to language. One feels the creature would just be blurting noises out, not speaking. We construct our sentences as we utter them (or a bit before), but these putative speakers are not constructing their utterances, just emitting them.  [2] We actively produce our utterances according to plan, but they passively emit noises when a stimulus strikes them. They can’t help speaking, but they can’t speak at will either. How could their utterances be productive if they don’t produce them—if they are extracted not created? The faculty of speech is essentially creative, given productivity, but then it can’t be elicited by stimuli in the way perception and reflexes are. Thus speech is necessarily stimulus-free. Active combination is its essence, but that is not compatible with the S-R model. We make utterances; they are not elicited from outside. And the same is true of thought for essentially the same reasons: thoughts are actively constructed complex entities not pre-existing fixed entities; so we can’t conceive of them as triggered in the S-R style. If thoughts were so triggered, they would not be thoughts, but something more like perceptions. What it is to be a thought involves stimulus freedom; this is not a contingent feature of thoughts. Thought, like speech, is essentially free—not an inescapable response to a exigent stimulus.  [3]

            We can therefore say that the form of sentences and thoughts is internally related to their being stimulus-free. If we call this form “grammar”, we can say that grammatical events are necessarily stimulus-free: what is composed in a certain way must come about in a certain way—freely, not by bondage to a stimulus. Form dictates etiology. Given that sentences have grammar, they must be stimulus-free; the price of stimulus bondage is loss of grammar. We can certainly imagine a machine that produces sounds like the sentences of a human language when stimulated to do so, but that is not to say that those sounds have grammar. In order to have grammar the sounds must be produced in a certain way (a derivational history), and that way excludes stimulus elicitation. The question is like asking whether genuinely creative acts could occur as responses to stimuli, and the answer is that these are incompatible attributes. Likewise, it is strange to suppose that a perceptual state could come about without a triggering stimulus—in the way that we speak and think. For that is incompatible with its defining attribute: a free-floating visual percept, produced at will, independent of any eliciting stimulus, would not be a genuine percept, but something more like a markedly visual thought. Perceptions must indicate the actual impinging environment and hence be responses to stimuli from that environment; they cannot be stimulus-free in the manner of language and thought. You change the nature of the thing if you change its mode of occurrence (or the laws of its occurrence). Language and thought, with their distinctive intrinsic structure, don’t just happen to be stimulus-free–as perception doesn’t just happen to be stimulus-bound. This distinction of mode of occurrence marks a deep ontological division in the mind. We might even say that we have two minds: an S-R mind and a non-S-R mind. We feel this division in ourselves all the time, as we register perceptions triggered by the environment and entertain thoughts stemming from who knows where. We consist of reflexes and spontaneities—bondage and freedom. We are slaves to stimuli and yet we are able to rise above them; we are not purely one thing or the other.

  [1] Readers who sense an affinity between what I say here and Fodor’s The Modularity of Mind are not mistaken. Encapsulation and reflexivity are connected properties.

  [2] Reflexive vocalization occurs in us too: we blurt out a vocal response to a stimulus, e.g. when injured and in pain. But these elicited vocalizations are not structured utterances, merely cries. This is not reflexively triggered speech.

  [3] I have not discussed other mental phenomena that raise the same question: dreaming, mental imagery, seeing-as, memory, knowledge of various kinds (perceptual, mathematical, moral, modal, introspective), logical reasoning, desire, emotion, intention, aesthetic response, and whatnot. About each of these we can ask whether they are stimulus-free or stimulus-bound (or possibly a mixture of the two), but I won’t attempt to answer this list of questions now–except to state my view that dreaming is free and memory is bound.