José Luis Bermúdez, The Paradox of Self-Consciousness. Cambridge, MA: MIT Press 1998, xv + 338pp.


Kent Bach

San Francisco State University



In many ways this is a masterful book. It is philosophy in the manner of Strawson, but richly fortified with experimental findings on a wide range of psychological phenomena. There is much to think about here and plenty to learn, as Bermúdez guides the reader on a well-organized tour along a progression of complex issues pertaining directly or indirectly to self-consciousness.

The goal of the book is to resolve its title’s paradox. Like most paradoxes, this one involves a set of seemingly true but mutually incompatible propositions. I won’t enumerate them here, but the upshot of this paradox is that an account of self-consciousness cannot avoid circularity. Now according to Bermúdez, giving such an account requires analyzing the capacity to think what he calls ‘I’-thoughts, as canonically expressed by means of the first-person pronoun (“analyzing” a capacity is not only to explain it but to characterize what constitutes it). But analyzing the capacity to think such thoughts requires analyzing the capacity to use the first-person pronoun, and that seems to require analyzing the capacity to think ‘I’-thoughts. How, then, to analyze self-consciousness without circularity (either explanatory or constitutive)?

As Bermúdez sees it, “the circularity at the heart of the paradox arises from the assumption that the capacity to think thoughts with the first-person contents characteristic of self-consciousness is available only to creatures who have mastered the semantics of the first-person pronoun” (p. 43). This assumption is a special case of what he calls the Thought-Language Principle:

Thought-Language Principle. The only way to analyze the capacity to think a particular range of thoughts is by analyzing the capacity for the canonical linguistic expression of those thoughts. (p. 13)

This principle may be broken down into two sub-principles, the Priority and the Conceptual Requirement principles. These have the effect, respectively, of tying conceptual abilities directly to linguistic abilities and of denying that nonconceptual states can have representational contents at all. To avoid circularity and thereby solve the paradox, Bermúdez will reject the Conceptual Requirement Principle.

In so doing he will rely heavily on a notion of nonconceptual content. “The central idea will be that both in explaining what mastery of the first-person pronoun actually is and in explaining how such a capacity can be acquired in the normal course of human development, we can appeal to nonconceptual first-person thoughts” (p. 45). Accordingly, Bermúdez defends the Autonomy Principle:

Autonomy Principle. It is possible for a creature to be in states with nonconceptual content, even though that creature possesses no concepts at all. (p. 61)

Without this principle the task of explaining self-consciousness (and cognitive development in general) would be faced with a dilemma: either we must deny that infants have the sorts of representational states that best explain the many surprisingly complex kinds of behavior they are demonstrably capable of, or we must ascribe to them mastery of concepts they could not possibly have. If we are to escape this dilemma and “do justice to both the differences and the similarities between infant and adult cognition then we will have to recognize the existence of states that represent the world in a way that is independent of concept mastery and, moreover, that can be ascribed to creatures who possess no concepts whatsoever” (p. 83).

Although Bermúdez is careful to distinguish constitutive from developmental issues, clearly he thinks that there is a connection between the two. In particular, he takes facts about cognitive development concerning the precursors of full-fledged self-consciousness to be strong evidence not just for ontogenetic but for constitutive claims. A major portion (chapters 5-9) of the book is devoted to delineating “a plausible developmental progression from the cognitive skills and abilities that normal human infants have available to them at birth via the relevant forms of nonconceptual self-consciousness to linguistic mastery of the first-person pronoun” (p. 269). In these chapters Bermúdez skillfully presents and applies an impressive range of recent psychological findings in order to trace the complex hierarchy of cognitive abilities that lead to full-fledged self-consciousness. Here I can only highlight some of the main elements of this progression.

Nonconceptual self-consciousness begins in infancy with somatic proprioception and with the pick-up of self-specifying information in exteroceptive perception. Although these comprise “a form of primitive self-consciousness operative in the very structure of perception” (p. 131), they involve “a relatively impoverished conception of the self associated with a comparably impoverished conception of the environment” (p. 272). Even so, visual perception and somatic proprioception are “the building-blocks for the bootstrapping process that will eventually result in the mastery of the first-person concept and the capacity for full-fledged self-consciousness” (p. 132), as well as in richer conceptions of the environment.

Somatic proprioception provides information about the state and position of the body at a particular location and about states of body parts relative to the rest of the body and to other body parts. Importantly, these are pieces of “self-specifying information.” So is the information provided by visual kinesthesis, which provides information on the perceiver’s movements, indicating that “the self has a place in the content of visual experience” (p. 112). Self-specifying information is also provided in how perceivers view objects in relation to their own action, e.g., “as within reach or as too heavy for one to lift” (p. 200). And the sense of touch, “because it is simultaneously proprioceptive and exteroceptive, provides an interface between the self and the nonself” (p. 164).

Bermúdez examines in detail how the child develops a nonconceptual point of view on the world, which “involves taking a particular route through the environment in such a way that one’s perception of the world is informed by an awareness that one is taking such a route” (p. 273). Bermúdez’s discussion (Chapters 7 and 8) of this and related phenomena, including the capacities to track perspectival changes in spatial relations among objects caused by one’s own movements and to think about places independently of what occupies them, is reminiscent of Strawson’s rendition of Kant’s metaphysics of experience in The Bounds of Sense (Methuen 1966), but is splendidly enhanced by the extensive research he invokes, e.g. on object permanence, autobiographical memory, spatial reasoning, and navigational abilities.

Finally, there are the abilities to think of oneself as a perceiver, as an agent, and as possessing various psychological properties. This last ability is, as Strawson argued long ago in Individuals (Methuen 1959), inseparable from the ability to ascribe such properties to others (it is essential to the capacity for “nonsolipsistic consciousness”), and that ability is essential to being able to draw others’ attention to things, to engage in joint attention and in joint activity, including communication.

Now it is time to register a few worries, none of which is meant to detract from the immense value of Bermúdez’s subtle and erudite account of the hierarchy of forms of self-consciousness.

(1) His paradox of self-consciousness depends on the Thought-Language Principle and, in particular, its two component principles:

Conceptual Requirement Principle. The range of contents that one may attribute to a creature is directly determined by the concepts that the creature possesses. (p. 41)

Priority Principle. Conceptual abilities are constitutively linked with linguistic abilities in such a way that conceptual abilities cannot be possessed by nonlinguistic creatures. (p. 42)

These two principles together comprise what Bermúdez calls the “classical view of content.” However, it is not obvious what is classical about this view or why he thinks it has been widely held—that needs documentation. I don’t doubt that there have been plenty of others who have held it, but the only philosopher he cites as holding it is Dummett, and Dummett’s views are not exactly classical. More importantly, Bermúdez does not explain why anyone should find the “classical” view plausible. To the extent that one does not, it is difficult to be exercised by his paradox.

Clearly Bermúdez himself finds the “classical” view plausible, at least plausible enough to be worth refuting. Indeed, he accepts the Priority Principle with little explanation or comment, noting that it “allows us to make a very clear distinction between conceptual and nonconceptual modes of content-bearing representation, because the connection between language and concepts gives us a clear criterion for identifying the presence of conceptual representation” (p. 43). But this doesn’t begin to show that there is such a connection, much less what that connection is. To suppose that linguistic abilities are necessary for conceptual abilities is to deny that even the most advanced apes possess concepts. Also, considering the extent to which language is rife with ambiguity and semantic underdetermination, it would seem that there are certain fine-grained conceptual abilities for which linguistic mastery is not sufficient. In any case, Bermúdez’s complaint is with the Conceptual Requirement Principle. Although he goes to sublime lengths to refute this principle, it is not clear why he finds it plausible in the first place. It seems that only a philosophical dinosaur, by maintaining that unconceptualized perception is mere sensation, would claim that seeing or feeling that p is impossible without thinking that p. It is hard to see or conceive why anyone in this day and age would insist that representational states must be conceptual.

(2) It is also unclear why Bermúdez takes the conceptual/nonconceptual distinction to pertain to content. For surely at least some conceptual states and some nonconceptual states have contents in common, e.g., the proposition that a certain tomato is red. But if states of both types can have identical contents, then how can the conceptual/nonconceptual distinction pertain to content itself? Fortunately, it is not essential to Bermúdez’s intricate resolution of his paradox that the operative distinction be one of content. Terminological adjustments aside, he could make all his main points while taking this distinction to pertain to types of states or perhaps to ways of taking contents.

My impression is that lurking in the background is a certain view of attitude ascription and a certain conception of propositional content. It is clear from what he says in various places (pp. 3, 33, and 83) that Bermúdez assumes that the ‘that’-clauses of attitude ascriptions fully specify or otherwise individuate attitude contents, but, as I have argued recently (Pac. Phil. Quar., Jan. 1997), this widespread assumption is problematic. He also seems to assume that the constituents of propositions expressed by ‘that’-clauses are concepts. In so doing, he effectively rules out the currently popular view that such propositions are “Russellian,” and have as their constituents not concepts but objects, properties, and relations. On that view, there is no meaningful distinction between conceptual and nonconceptual content. If attitudes and experiences have propositional contents of this sort, then obviously the conceptual/nonconceptual distinction can meaningfully apply only to states that have them or to ways of taking them.

(3) Bermúdez does not set the problem of explaining self-consciousness against a background account of what is involved in being conscious of objects in general. Nor does he set the problem of explaining mastery of the first-person against a background account of what is involved in mastery of singular terms in general. It would be unreasonable to expect a general treatment of singular thought and singular reference in an already long and involved book, but it would have been desirable to have an explanation what is special about thinking and referring to oneself, as opposed to anything else. Who knows, perhaps an adequate account of what is required for consciousness of and reference to things in general would show that nothing special is required for self-consciousness and self-reference. Indeed, it seems to me, such an outcome would be consonant with Bermúdez’s many observations about the complementarity of object- and self-consciousness.

(4) It is puzzling why Bermúdez finds it even initially plausible that the capacity for self-consciousness should be intimately tied to the capacity to use the first-person pronoun. Particularly puzzling is why, in the discussion which occurs almost as an afterthought in the concluding chapter, he thinks that the communicative use of ‘I’ (its use to refer others to oneself) is essential to the analysis of the capacity for self-consciousness. And, although he rightfully argues that there are the forms of self-consciousness which do not require the capacity to use the first-person pronoun, there are also aspects of full-fledged self-consciousness which require more than linguistic mastery of the first-person pronoun. He acknowledges that “forms of self-consciousness [vary with] the richness of the conception of the self that they provide” (p. 272), but the richer ones can and do go far beyond what is required for competent use of the word ‘I’. They pertain to what one takes oneself to be, and it is not possible to take oneself to be anything without already being able to think of oneself.

Bermúdez plausibly holds that “a crucial element in any form of self-consciousness … enables the self-conscious subject to distinguish between self and environment” (p. 272). Several of the cognitive agnosias he considers illustrate how this distinction can become distorted in one way or another. I wish he had looked into some of the more extreme psychiatric ways in which this distinction can break down. People with chronic uncontrollable impulses, radical mood swings, multiple personalities, or schizophrenic breaks have to different degrees difficulty in forming a stable self-conception, but their problems go well beyond anything having to do with mastery of the first-person pronoun. In extreme cases, a person’s sense of where the boundary is between himself and the rest of the world is not at his skin but somewhere well within it, perhaps not far from the pineal gland.

* * * * *

Even though the paradox of self-consciousness rests on principles that I find much less plausible than Bermúdez does, I am glad he was gripped by it. This enabled him to produce a rich and deeply rewarding book on one of the most difficult topics in philosophy. No philosopher heretofore has come close to bringing such a wide range of scientific findings to bear on self-consciousness in its many stages and aspects. Paradox or no paradox, the reader can safely venture into the Bermúdez triangle. An edifying experience awaits.