Delivered at Stanford University in February 1980, Quine’s Immanuel Kant Lectures titled Science and Sensibilia have, until now, remained unpublished.¹ Taken as a whole these lectures fill an important gap in understanding the development of Quine’s later thought, especially as they mark the middle point between his earlier The Roots of Reference (1974) and his later Pursuit of Truth (1992). In general, the lectures further develop Quine’s physicalistic program of tracing the links from sensory stimulation to our scientific speculations about the world, better known as his naturalization of epistemology, but it does so by framing the discussion in terms of two further, more ‘traditional sounding’, sub-questions in epistemology. These questions concern our knowledge of the external world and our knowledge of other minds. Obviously, this familiar philosophical rendering of such questions will need to be modified in order to fit the standards appropriate to the physicalist perspective found in these lectures. In Quine’s hands such questions are transformed into a concern over how on the basis of sensory stimulation we have come to make statements about the world and other minds. But Quine’s willingness to use these questions within his lectures suggests a somewhat more concerted effort on his part to highlight how his interests remain connected to the traditional questions of epistemology. Even when, as we will see, his discussion significantly departs from these more familiar formulations and in the methods used to address them (Fig. 1.1).

The lectures begin with Quine’s own familiar rendering of the epistemological question in this slightly new setting: “How, on the strength of the mere sporadic triggering of our sensory receptors, is it possible to fabricate our elaborate theory of other minds and the external world?” (Quine, This volume, 19). With this as his overarching concern, Lecture I lays out the basic framework within which he plans to address this question. He begins by briefly arguing for the merits of a monistic physicalism and wonders to what extent our mental talk, or “mentalistic” vocabulary, can be captured within, or adapted to such a physicalist framework. He gives the following standard for physicalist adequacy: “A supposed mental state or event qualifies as physically genuine if it is specifiable strictly by physiological description, presumably neurological, without recourse to mentalistic terms” (Quine, This volume, 24). Quine then extends this discussion to perception, where he assesses the physicalist credentials of ‘perceptual events’ by arguing that they can all be grouped under a neurological formulation (Quine, This volume, 25–31). Moreover, because the subjective similarity between perceptual events yields similar behavioral responses, it is assumed that similar perceptual events will also be similar in their neural mechanisms (Quine, This volume, 27).

With this physicalist rendering of perception in place, Quine begins to discuss the route from perception to scientific theory, where this turns on an examination of the acquisition of cognitive language. Lecture II continues this account tracing the steps from the learning of observation sentences through ostension to standing sentences and lastly predication and relative clauses. With Lecture III Quine directs his attention to the second sub-question mentioned above concerning our knowledge of other minds. Here, we learn that observation sentences, sentences that are primarily about features of the external, physical world and not sensations, rather easily yield mentalistic counterparts that are observation sentences about other minds. Quine then explores how to make sense of such sentences within his physicalist framework. Of special interest is how he attempts to make physicalist sense of the objects of perception, a topic that I will return to below. Lecture IV then concludes with Quine’s reflections on ontology where we see him moving closer to his mature position that emphasizes the significant role of structure rather than objects for our ontology.

Much of the lectures are then aimed at locating mentalistic discourse within a scientific, physicalist framework, where this is offered as a scientific, if abstract explanation of how we come to know the external world and other minds, without an appeal to mental entities or other sensibilia. Quine mentions at the outset his gratitude for the resounding title of his lectures, Science and Sensibilia, to John Austin’s take on Jane Austin (Quine, This volume, 20). Here, of course he is referring to Austin’s Sense and Sensibilia, which famously criticized the concept of ‘sense-data’ and the so-called argument from illusion. A somewhat closer look at Austin’s criticisms will, I think, be useful for understanding why Quine adopted this title for his lectures, while also serving to help clarify his approach in these lectures.

Austin’s focus on ordinary language analysis is central for his attempt to eliminate the philosophical errors introduced by the misuse of words and mistaken accounts of the features of the world picked out by descriptive terms. The careful regard for the way expressions are used in everyday language is presented both as a necessary prelude to guard against philosophical error and for providing a proper account of the phenomena under consideration (Berdini and Bianchi 2018; Longworth 2017, 2.1).

This method is on full display in Sense and Sensibilia, where Austin seeks to discredit the general view that claims “we never see or otherwise perceive (or ‘sense’), or anyhow we never directly perceive or sense, material objects (or material things), but only sense-data (or our own ideas, impressions, sense, sense-perceptions, percepts, etc.).” (1962, 2). Much of the supposed support for such a position stems from what is often called the argument from illusion. In cases of illusion we have a sensory experience of seeing something with certain specific traits or features, but where nothing has these specific traits. This might be because the experienced object in question lacks these features, or that we simply do not experience such an object whatsoever. Now in such experiences it is often thought that there must be something that has these experienced features, usually called sense-data. Because in such cases there are no material objects of the required sort experienced, sense-data are themselves not material objects, nor parts of the surrounding environment that are independent of our individual experiences. What are then experienced directly are these sense-data, which themselves are taken to be distinct from material objects. A further inference is now made to the conclusion that every experience (whether illusory or not) has a similar character, so our experience of material things always has sense-data as its object. It is never the case that we directly experience material things.²

Consider the classic case of refraction. The stick which normally looks straight appears bent when seen in water. But the stick does not change shape when it is placed in water and therefore it cannot be both bent and straight. One of these visual appearances must then be delusive, but even when what is seen is not a real feature of an object, it is still the case that we see something. And it is this sense-datum that we are taken to be directly aware of in sensory perception (Austin 1962, 21). Or consider the case of mirror-images, that is, mirror reflections. When someone sees themselves in the mirror, their body appears to be some distance behind the glass, but they cannot actually be in two places at once. Once again, these perceptions cannot all be accurate, so our image is not in the place it seems to be. We must be perceiving something else, which is a sense-datum.

In addressing this view Austin begins by emphasizing that there is no single way in which we may be ‘deceived by the senses’ (that is, perceive something unreal or not material), and further that “things may go wrong […] in lots of different ways—which don’t have to be, and must not be assumed to be, classifiable in any general fashion” (Austin 1962, 13). He further wonders if we tend to speak of ‘illusions’ with reference to dreams, photos, mirror-images or pictures seen on the cinema screen. By pointing out the familiarity of the circumstances where we meet these phenomena and the ordinary methods used to consider them, Austin aims to show how the distinction between sense-data and material objects, and between illusory perceptions and veridical ones, offers us false alternatives, when compared with our prior ordinary methods for dealing with ‘illusions’ (Berdini and Bianchi 2018).

We can further expand on this idea by considering an additional element of Austin’s response to the sense-data theorist. Here he makes appeal to two components present within our ordinary perceptual judgments involving what has been described as the “opportunity afforded by sensory perception” and our general expertise and abilities in forging appropriate judgments given those opportunities (Longworth 2017, 3.2). This further enables him to clearly distinguish the difference between illusion and delusion and how this helps to explain the above cases without any appeal to sense-data. In the case of the bent stick, the sense-data theorist claims that the false perceptual judgment is dictated by the sensory experience, which looks as if the stick is bent. But because it isn’t really bent, what we experience must be the relevant sense-data. Austin places special emphasis on our ordinary ability to make perceptual judgments, thereby denying sensory experience by itself must dictate our perceptual judgments (1962, 29–30). Consider this basic fact: A stick when placed in water looks bent, but it doesn’t look like a bent stick when not in water. A key question here concerns whether it is the case that something that is straight must always look straight? Our ordinary perceptual practice suggests not, since we willingly admit that the same thing can look different depending on a variety of different possible circumstances. The idea that we must see something differ...