In this book Tyler Burge brings together ten papers on Frege that he published from 1979 to 2000. The volume contains more recent postscripts to four of the essays and a substantial new introductory essay just short of seventy pages in length. All but one of the essays are grouped in three parts dealing in turn with ‘Truth, Structure, and Method’, ‘Sense and Cognitive Value’, and ‘Rationalism’, though, as one might expect, there are themes that extend through more than one part. A short essay outlining Frege's philosophical ideas is placed between the Introduction and Part I.

These are not easy pieces but dense, challenging essays directed towards teasing out ideas that are fundamental to Frege's thinking about logic, language and the foundations of arithmetic. Burge gives us a glimpse of what drives him to devote such care and attention to close reading of Frege's texts when, in the Introduction, he writes,

Burge explicitly aims to avoid assimilating Frege's preoccupations to those that prevailed after his time, for instance, among the logical positivists. He also wants to steer clear of the kind of criticism that fails to have due regard for Frege's detailed aims and background assumptions. He offers us, not an easily accessible narrative presenting a cleaned up, sorted out version of Fregean thought, but a series of close scholarly studies. He provides copious textual evidence to back up his interpretations and enable readers to check his attributions for themselves.

Burge (ch. 6) defends the view that for Frege ‘expressions of mathematics, science and ordinary discourse are not sharply understood, even by their most competent users’ (p. 251). That looks puzzling. How can expressions used in doing mathematics have a definite sense that even specialists in mathematics do not grasp? This takes us to a key theme of this body of work: ‘for Frege the capacity of a word to express a sense is partly independent of the user's understanding and use of the word, and even of conventions about the word's usage, in a way that distinguishes his notion of sense from almost any ordinary notion of linguistic meaning—conventional or idiolectic’ (p. 255). How can this be? Burge acknowledges that Frege does not directly confront the issue. But he discerns in Frege's writings the following position.

This is far more radical than the thesis that one may count as, for instance, meaning arthritis by ‘arthritis’ even though one's dispositions to use the term embody a partial, or even partially erroneous, grasp of what it means. The latter thesis is compatible with the claim that some people—the experts—must have full mastery of the term's meaning and of the concept of arthritis. But, on the view attributed to Frege, the sense of an expression might not be fully grasped even by the most competent users. Yet in glossing the position Burge retains a connection between the sense, and thus the denotation, of an expression and the practice to which its use relates. He says, ‘The implications of a practice can reach beyond the procedures and dispositions of the practitioners’ (p. 258). Many of our practices have a deep rationale. For Frege, Burge suggests, the rationale that underlies some of our practices ‘makes its imprint on those practices in such a way that understanding the rationale is achieving insight into what was part of the practices even before the understanding was achieved’ (p. 258). This intriguing idea seems to me not only to be important for the interpretation of Frege but also to merit investigation for its own sake. An investigation of this sort might well have ramifications for our understanding of what philosophy is about, particularly when conceived as a form of conceptual enquiry.

Another key theme is that although Frege does not discuss epistemology in general, or even say much in detail about the theory of knowledge of arithmetic, epistemological considerations inform his thinking about logic and mathematics. Burge presents Frege as having ‘accepted the traditional rationalist account of knowledge of the relevant primitive truths, truths of logic’ (p. 299). But, especially in chapter 9, he is at pains to stress that although Frege took primitive truths to be self-evident, he did not conceive of self-evidence in psychologistic terms. Primitive truths need not be obvious; their self-evidence would emerge only if one obtained a proper understanding of them—an understanding that can elude even specialists. (See the treatments of fallibilism in section IV of ch. 9.) There is an obvious link with the idea about sense described above. Self-evidence belongs at the level of the deep understanding that Frege hoped to achieve. Obviousness may reflect a partial or defective understanding that informs our actual thinking prior to having reached the deep understanding.

One of the essays (ch. 8) interestingly explores the relation between Frege's conception of the status of primitive logical truths, conceived as laws of truth, and considerations about what is reasonable. Frege says that ‘from the laws of truth there follow prescriptions about asserting, thinking, judging, inferring’(p. 313). The order of explanation, Burge observes, is paradigmatically Platonist: ‘from what is in an abstract realm to what is reasonable’ (p. 313). The link is forged by the idea that judgement of its nature aims at truth. So the prescriptions that are derivative from the laws of truth ‘must apply to the judgements of judging subjects’ (p. 314). Against this background Burge argues that the conventional objection to Platonism about mathematics is misconceived. The objection is that it's a mystery how we could have the causal commerce with an abstract realm that is supposed to be necessary for knowledge of a mind-independent reality. For Frege the issue of causal commerce does not arise because ‘to be a judging subject is to be subject to the prescriptions of reason, which in turn are provided by the laws of truth (logic)’ (p. 316). Instead of focusing on how an abstract realm could impinge upon us we should look to what the primitive truths require of us given that it belongs to us as judging subjects to aim at truth. Closely related matters are explored in an essay on Frege's understanding of the a priori (ch. 10). Here Burge resists Frege's view that apriority is to be conceived in terms of generality while taking seriously the idea that ‘reason … [is] essentially involved in supplying general principles and rules of inference’ (p. 373). This theme is pursued by tracing points of kinship and contrast between Frege and Kant on knowledge of arithmetic and knowledge of geometry.

A long essay from 1986 on Frege on truth (ch. 3) and an extensive postscript from 2004 nicely illustrate Burge historical approach to understanding Frege. Burge rejects Frege's view that declarative sentences denote truth-values—true sentences the True, and false sentence the False. He aims to show that this view ‘has profound and natural motivations’ (p. 84), and that the associated thesis that truth-values are objects ‘is more pragmatically based—and therefore less strange—than has usually been thought’ (p. 84). A crucial consideration here is that ‘word denotation is important because and only because of the importance of some feature of sentences that is central to logical theory and functionally dependent on word denotation’ (p. 87). Granted that the truth-value of a sentence is dependent on the denotations of its constituent expressions, why must we think of the truth-value as an object that is the denotation of the sentence? Burge's answer is that Frege was exploiting natural analogies between the logical roles of whole sentences and the logical roles of singular terms in a manner that was theoretically convenient. ‘The places of sentences, like the places for terms, stand in argument places for functional expressions; they sometimes constitute value expressions resulting from functional application; and they never stand for functional or predicate expressions’ (p. 98). The postscript to the essay on truth adds a discussion of logical validity and logical consequence and engages critically with the interpretative work of Thomas Ricketts.

Taken as whole the essays address all of the central ideas and problems in Frege's philosophical work. There are, for instance, extensive discussions of Logical Law V, of the extensions of concepts, and of the hierarchy that is generated on Frege's conception of how to deal with apparent failures of substitutivity. This is a rich and immensely impressive body of work, which will be indispensable reading for specialists, but contains much that may stimulate and provoke those who think about the topics covered for their own sake.