|
in the history of analytic philosophy *
Though the genre has existed since the middle of the twentieth century, reflection on analytic philosophy’s history and nature has come into its own as a field of inquiry only in the last fifteen years. In this essay, I aim, first, to survey some of the main ways in which analytic philosophy has been represented in historical work concerning it, and second, to discuss the implications of certain inconsistencies in these representations. Specifically, I will argue that (1) historical work on analytic philosophy has undergone an evolution that can be parsed roughly into three main phases, (2) work in the first phase helped to solidify a received view of analytic philosophy which, though erroneous, was crucial to the movement’s success, (3) work in the second and third phases have a shared tendency to define analytic philosophy in ways that fail to do justice to the facts surrounding the rise of analytic philosophy, especially as concerns the historical significance of the erroneous received view, and (4) all of this points to the conclusion that analytic philosophy is something of an illusion. the 3 stages of historical work on analytic philosophy On my view, historical work on analytic philosophy falls roughly into three categories, which I prefer to characterize as three evolutionary stages of its development.1 The first stage consists of contemporaneous first-hand accounts of analytic philosophy in its early and middle stages (approximately 1900-1950), and near-contemporaneous, memoir-like accounts of the same. Gilbert Ryle, himself an important contributor to this category, once observed that “history begins only when memory's dust has settled” (Ryle 1963, 1). The defining characteristic of work in this category is that it came into being while memory’s dust was still loose in the air: the “analytic philosophy” of which its authors speak is, for them, either a living reality or at least one retained in living memory; the observations are largely first-hand even if retrospective, as opposed to current work which reconstructs the early history of analytic philosophy on the basis of written records. Given its memoir-like status, we may call work in this category “proto-history”. Proto-historical works on analytic philosophy include J.O. Urmson’s 1956 Philosophical Analysis, G.J. Warnock’s 1958 English Philosophy Since 1900, and the 1963 collection The Revolution in Philosophy, which includes essays by such analytic luminaries as A. J. Ayer, P. F. Strawson, and Gilbert Ryle.2 One notable feature of work in this proto-historical stage is the tendency of its authors—especially those writing from within the analytic tradition—to act as if the “essence” of analytic philosophy was relatively easy to pick out. They seem to take it for granted that analytic philosophy originated around the turn of the twentieth century in a radical break with philosophy in the great tradition, and that this revolution in philosophy was founded primarily on a novel method—the analysis of language—and the novel metaphilosophical view it inspired, namely, that philosophers should conduct their business by analyzing language because analyzing language is the only business that philosophy can legitimately claim as its own. The tendency among proto-historical authors is to treat this metaphilosophical view as the central doctrine of analytic philosophy. They take it for granted that analytic philosophy’s other characteristic features, such as its anti-historical and anti-metaphysical tendencies, all derive from its core belief that the right way to do philosophy had been discovered, and that it was the analysis of language. Finally, they take it for granted that the main figures responsible for the revolution were, first, Moore and Russell (and Frege, insofar as Russell appropriated his techniques in mathematical logic), and, later, Wittgenstein. Works about analytic philosophy written during this early period were guided largely by this received view. It is true that they frequently mention the fact that there were differences among analytic philosophers, and they note the confusion this sometimes caused for people (usually characterized as “outsiders”) trying to understand the movement. However, in the usual case, these differences are quickly dismissed as inconsequential in light of the deeper unity to be found in the analysts’ common acceptance of the analysis of language as the sole legitimate, or at least the most important, mode of philosophical activity. For example, Arthur Pap claimed that, though there were significant differences among analytic philosophers, “the unanimous practice of the analytic method as a powerful instrument of criticism tends to blur these differences…” (Pap 1949, ix). What was this unanimously-practiced analytic method? Pap preferred to call it ‘logical analysis’, but it is clear that he had in mind the analysis of language, broadly construed: in general, all the typically philosophical questions of the form ‘what is the nature of X’ can be interpreted as questions of logical analysis, of the form ‘what is the meaning of the word ‘X’ or of any synonym thereof,’ or ‘what is the meaning of sentences containing the word ‘X’. (Pap 1949, vii).Thus, for Pap, analytic philosophers were united in their practice of the analytic method understood as the analysis of language. Similarly, though Urmson notes that “the analytic practice had no clearly defined dogmatic background at all” concerning the objects of philosophical analysis, he goes on to say that analytic philosophers were united at least in the view that analysis was at least one of the most important tasks of the philosopher; and by analysis, they meant something which, whatever precise description of it they chose, at least involved the attempt to rewrite in different and in some way more appropriate terms those statements which they found philosophically puzzling. (Urmson 1956, vii)By emphasizing its linguistic aspects while simultaneously diminishing the significance of any theoretical disagreements among analytic philosophers that might have undermined the linguistic interpretation of philosophical analysis, Urmson’s description reveals the centrality of that interpretation for the mid-century conception of analytic philosophy. This pattern of emphasis and diminution is even clearer in a similar statement from Peter Strawson’s 1963 essay “Construction and Analysis”. Strawson was aware that there was considerable divergence of opinion over just what the objects of philosophical analyses were, even noting that this seemed to have an effect on the kind of enterprise a philosopher is engaged in—if they are sentences or statements, then philosophy is like grammar or linguistics, if they are thoughts or beliefs, then philosophy is like psychology. Nonetheless, Strawson affirms that the unity of analytic philosophy is grounded in the unanimous practice of the analysis of language: It does not matter much … [what we say the objects of analysis are], … Maybe it is best to say, as Moore always said, that the objects of analysis were propositions. This answer, whatever its shortcomings, emphasizes, without over-emphasizing, the linguistic nature of the enterprise, the preoccupation with meaning. For, however we describe the objects of analysis, particular analyses … always looked much the same. A sentence, representative of a class of sentences belonging to the same topic, was supposed to be elucidated by the framing of another sentence. (Strawson 1963, 98; my emphasis)Again we see an insistence upon the view that philosophical analysis is the analysis of language; any theoretical differences that might have suggested otherwise are shrugged off as unimportant relative to overwhelming similarity in the linguistic aspects of the practice of analysis. In sum, then, what we see coming out of the proto-historical stage of historical work on analytic philosophy is a record of what was then, and what by and large has continued to be, the received view of analytic philosophy. On this received view, analytic philosophy is a school of philosophy that originated in a revolutionary break with philosophy-in-the-great-tradition around the turn of the twentieth century. The break was fueled by the perception that the correct method of philosophical inquiry had finally been discovered, and that it was the analysis of language (hereafter, I shall call this view the linguistic thesis). The second half of the twentieth-century saw astonishing changes in the analytic world, changes that would ultimately make it impossible for the received view to persist. In the 1960s and 70s, analytic philosophy’s linguistic character began to fall away, and metaphysics reemerged as a legitimate enterprise. In the 1970s and 80s, analytic philosophy’s anti-historical attitude began to loosen up, and space was made within the social scope of the movement for people to do more purely historical work on the history of philosophy (cf. Schneewind (ed.) 2004). By the early 1990s, this new historical approach was adopted by philosophers interested in applying it to the history of analytic philosophy itself. Thus emerged a second stage in the historiography of analytic philosophy, which I call “new wave” history. New wave history is exemplified by such figures as Nicholas Griffin, Peter Hacker, Ray Monk, Peter Hylton, and Michael Beaney, among others. The title “new wave” signifies not only the use of the new historical approach, but also the fact that the results of their studies frequently challenge the received view of the proto-historical period, which persists today in a somewhat altered form, expanded so as to accommodate the developments within analytic philosophy during the latter half of the twentieth century—it is our received view (cf. Preston 2004, 2005). In my estimation, the most important finding of new wave scholarship is that no view traditionally connected with analytic philosophy was actually shared by all and only canonical analysts—not even the linguistic thesis, which, as we have seen, seemed absolutely central to the analytic self-image in the proto-historical period.3 This lack of a common view does not result merely or even primarily from the fact that more recent analytic philosophers have abandoned the views of earlier analytic philosophers, so that, as Richard Rorty has observed, “most of those who call themselves ‘analytic philosophers’ would now reject the epithet ‘linguistic philosophers’ and would not describe themselves as ‘applying linguistic methods’” (Rorty 1992, 374 n. 9). Rather, the deeper problem, which the new wave scholars have rooted out, is that the traditional defining doctrines of analytic philosophy never achieved universal acceptance even among core, canonical analysts in the early and middle periods of the movement. By reading the main works of canonical analytic figures in light of each one’s broader corpus and their respective intellectual and social contexts, the new wavers have discovered what some of the proto-historians already knew: there were deep differences among the central movers and shakers in early analytic philosophy over issues so fundamental as what philosophical analysis was, and what the objects of analysis were. At least one canonical analyst, G.E. Moore, did not conceive of the objects of philosophical analysis as linguistic at all.4 Others did—such as Wittgenstein (in both his early and late phases), the Logical Positivists, and the Oxford ‘ordinary language’ philosophers—but they had different conceptions of what it meant to be a linguistic entity, what language was, how it functioned (i.e., how it meant, referred, etc.), and what the significance of all this might be for philosophy-at-large.5 Though this has the status of newly acquired knowledge for many of the new wavers and their generation, those in the proto-historical period were not unaware of these differences; as we have seen, they acknowledged them quite explicitly. The crucial difference is that, whereas the proto-historians saw these differences as trivial, at least some of the new wavers have taken them to be significant enough to undermine the received view and to send them searching for new ways of conceptualizing analytic philosophy. I will have more to say about this shortly, but we must first turn to the third stage of development in historical writing about analytic philosophy. We may characterize work of the third stage or type as “analytic history” for several reasons. First, it tends to be written by philosophers who work mainly in what are now called “core analytic” areas—philosophy of language, metaphysics, and epistemology—“hard-core” analytic philosophers we might call them, or to borrow John Ongley’s clever term, “high church” analytic philosophers. As one might expect from people who have their understanding of analytic philosophy as well as their general intellectual habits formed in these contexts, they exemplify a mentality and a method very different from the new wavers. In fact, it is much closer to what was standard in analytic circles prior to the historical movement of the 1970s. At that time, if the history of philosophy was studied at all, it was studied in the form of rational reconstructions of the views of historical figures, usually taken out of context and anachronistically assimilated to current interests and approaches. Thus, a second reason to call this “analytic history” is that it can be characterized as an application of the traditional analytic approach to the history of philosophy—the one against which the historical movement rebelled—to the history of analytic philosophy itself. What is most characteristic of analytic history is a tendency to work within the parameters of the received view, in some cases despite the fact that it has been severely shaken by the findings of new wave history. An early, paradigmatic case of analytic history is Michael Dummett’s Origins of Analytic Philosophy (Dummett 1993). The influence of the received view can be discerned in his choice to define analytic philosophy in terms of a metaphilosophical view involving the analysis of language: What distinguishes analytic philosophy, in its diverse manifesttations, from other schools is the belief, first, that a philosophical account of thought can be obtained through a philosophical account of language, and, secondly, that a comprehensive account can only be so obtained. (Dummett 1993, 4 f.)Or, as he also puts it, the “fundamental axiom of analytical philosophy [is] that the only route to the analysis of thought goes through the analysis of language.” (Dummett 1993, 128) Dummett’s book was written after the historical movement but largely before the new wavers arrived on the scene. When they did, they quickly made Dummett’s definition their whipping-boy. For example, Ray Monk has argued that, on Dummett’s characterization, even Bertrand Russell fails to qualify as an analytic philosopher (Monk 1997). Insofar as Russell is widely considered to be a patriarch of analytic philosophy, Monk’s argument amounts to a reductio ad absurdum of Dummett’s definition. A similar argument could be made putting G. E. Moore in the place of Russell. Beyond this, Dummett’s interpretation of Frege has been challenged (Hacker 1997, 52 f.; cf. Baker & Hacker 1983, 1984, 1987, 1989), so that perhaps even Frege fails to meet the criteria Dummett purports to draw from Frege’s own work. Given the utter untenability of Dummett’s definition in light of the historical and textual facts, where, we may wonder, did he get the idea for it? The only reasonable answer, it seems to me, is that he was guided by the received view. A similar influence is discernable in what is arguably the best and most important example of analytic history to date: Scott Soames’ monumental Philosophical Analysis in the Twentieth Century (Soames 2003). Unlike Dummett, Soames had the advantage of writing after new wave history had begun to make a noticeable mark.6 Consistent with new wave findings, Soames eschews a doctrinal definition of analytic philosophy, instead characterizing it as a “trail of influence” beginning with Moore, Russell, and Wittgenstein (Soames 2003a, xii f.). But this cannot really be Soames’ conception of analytic philosophy, for it cannot justify the principled selectivity Soames exhibits in tracing what he presents as the central storyline of analytic philosophy’s development. Trails of influence can easily be traced beyond the canonical domain of analytic phiosophy, whether we understand that in historical or thematic terms. For instance, Frege and Peano influenced Russell in ways that helped him make significant strides toward Principia Mathematica. And yet, neither of them is included as a key player in Soames’ history—Frege is merely mentioned several times, and Peano not at all. Either these limitations are arbitrary, or Soames’ “trail of influence” is circumscribed by something more substantial, by a different conception of analytic philosophy according to which something other than influence holds it together and sets its boundaries. And here, again, I suggest that the best candidate is the received view, for Soames focuses on just the figures and exactly the issues it designates as being central to the rise and development of analytic philosophy. Indeed, Soames comes close to acknowledging the received view in at least one place. By neglecting Frege, Soames is aware that he is leaving “an undeniable gap in the story” of analytic philosophy (Soames 2004b, 462). However, he excuses this on the ground that most of Frege’s work, which was done in the last quarter of the nineteenth century, “falls outside our official period” (Soames 2004b, 461). Given the scope of his history, it is clear that Soames’ official period begins around the turn of the twentieth century, with the work of G.E. Moore—exactly as the received view has it. current definitions of analytic philosophy So much, then, for the evolution of historical work on analytic philosophy. What has come of it? Despite the flourishing of work in this genre, little light has been shed on analytic philosophy’s true nature, or on the historical and intellectual factors most responsible for its advent and quickly-gained ascendancy in certain circles of academic philosophy. In fact, much current work in this genre is driving us further away from achieving this kind of understanding. This is because the inadequacy of traditional definitions of analytic philosophy, brought to light by the new wavers, has driven philosophical historians of all stripes to suggest a host of new definitions which diverge in significant ways from the received view. These definitions differ not only in respect of the particular features proposed as definitive of analytic philosophy, but also in respect of the types of features proposed as definitive. Hans-Johann Glock (Glock 2004) has recently provided a helpful taxonomy of the definitional types currently in circulation. His categories are:
1. doctrinal (in terms of the views analytic philosophers espouse) Glock’s taxonomy adequately captures the range of definitional types currently in circulation. However, most are flawed in ways that warrant their rejection. Some authors, including Glock himself, have marshaled arguments against some one or more of these definitional types, or, more frequently, against particular cases falling under them. Important as many of these arguments are, I am going to ignore them in order to focus on a general argument against all but the doctrinal type of definition. My claim is that all the other types fail qua types precisely because their generic content is not of the right sort to pick out a group of the sort usually called a philosophical school, movement, or tradition. In what follows, I will use these terms synonymously, and will frequently use “school” to stand for them all. It seems to have gone completely overlooked in the current definitional controversy over analytic philosophy that the way we define a philosophical school has metaphilosophical implications. There are different kinds of definition, but the kind traditionally aimed for in philosophy—and the kind needed in the case of analytic philosophy—is called “real definition”. Real definitions are supposed to pick a thing out according to its most fundamental, or essential, features, its necessary and sufficient conditions. With the exception of those in the family resemblance category, most other current attempts at defining analytic philosophy seem to be attempts at real definition—at least their proponents make no effort to explain that they are trying to provide something different. Thus, in proposing to define analytic philosophy topically, methodologically, or stylistically, genetically, or however, one implicitly proposes that topics, methods, styles, lineage, or whatever, are most fundamental to something’s nature as a philosophical school. But this is at variance with what most philosophers, both currently and historically, take to be true of the philosophical schools with which they affiliate. With the exception of a few dyed-in-the-wool Wittgensteinians of a certain variety, few in the contemporary analytic world would deny that philosophy is a theoretical discipline. Its business is, minimally, the production and critical assessment of theories by means of reasoning. Theories, minimally, are sets of views (propositions) about the way things are, or what is the case, in some region or other—or possibly the whole—of reality. And, again minimally, in order for philosophers to deal with such views corporately, they must be verbally articulated in a relatively straightforward way, in the form of a sufficiently clear declarative sentence. I trust it will be recognized that this minimal conception of what philosophy is and what it involves has been widely held, at least implicitly, throughout the history of the discipline. Now, this minimal metaphilosophical view has implications for how the emergent social world of philosophers (academic or otherwise) ought to take shape. On this view, what is most fundamental to philosophy is reasoning, on the one hand, and the objects of reasoning—ideas, views, and so forth—on the other. The constant in this pair is reason, and the variables are the particular ideas or views to which reason is applied. Thus, insofar as there are philosophically relevant divisions to be made within the social world of philosophy, they will be made along ideological lines. This suggests that there is a minimum standard, a necessary condition, for the initial formation and the retrospective demarcation of groups that, like schools, movements or traditions, purport to mark out not merely a region of social space, but of philosophical space: such groups must rely for their cohesion, and hence also their existence, on a kind of unity that is constituted by agreement in theoretical matters. That is, a group is most properly called a philosophical school (etc.) only when it has come together on the basis of a shared philosophical view (or some set of them). With this in mind, I shall say that a group counts as philosophical in the most proper, primary, or focal sense if and only if its criterion for membership is acceptance of some set of views on the basis of rational understanding. I will say of any group which meets this requirement that it is philosophically unified, or that it possesses philosophical unity. And, when a view actually functions in this way to ground the unity of a group, I shall call it a defining doctrine of that group. What is fundamental, then, to the sort of group commonly called a philosophical school, is its defining doctrines. And, since a definition is supposed to pick something out according to its most fundamental features, the doctrinal approach to definition is the only legitimate one for a philosophical school. Looking to styles, topics, methods, or anything other than defining doctrines is either to mistake the accidental for the essential or to misunderstand the nature of a philosophical school. an evaluative taxonomy The position described in Section 2 provides the foundation for a larger taxonomy that runs somewhat skew to Glock’s:
The categories in this taxonomy are not merely descriptive but evaluative; for, as I shall argue, falling into one of the left-hand options at any fork marks a definition as defective. The remainder of this essay will be given to explaining and systematically rejecting each of those left-hand categories until the illusionist variety is the only option left on the table. In fact, we have already taken a first step toward this conclusion. The first division in this new taxonomy is between doctrinal and non-doctrinal definitions. A doctrinal definition is one framed in terms of a school’s defining doctrines, in accordance with the position laid out in Section 2. A non-doctrinal definition is one framed in terms of anything else. We have already seen an example of a non-doctrinal definition in Soames’ characterization of analytic philosophy as a trail of influence. Here are two other examples. Brian Leiter gives a stylistic definition, saying: ‘Analytic’ philosophy today names a style of doing philosophy, not a philosophical program or a set of substantive views. Analytic philosophers, crudely speaking, aim for argumentative clarity and precision; draw freely on the tools of logic; and often identify, professionally and intellectually, more closely with the sciences and mathematics, than with the humanities. (Leiter 2000).Avrum Stroll gives what is perhaps best characterized as a family-resemblance definition: “it is difficult to give a precise definition of ‘analytic philosophy’ since it is not so much a specific doctrine as a loose concatenation of approaches to problems” (Stroll 2000, 5). According to the position of Section 2, non-doctrinal definitions are inadequate to define a philosophical school and should be rejected. Focusing only on the doctrinal definitions currently in circulation, we can make a further division: some doctrinal definitions of analytic philosophy are traditional, others revisionist. Traditional definitions are doctrinal definitions that keep to the received view, such as Dummett’s. Other, more recent characterizations fit the received view just as well. For instance, John Searle describes analytic philosophy as the dominant school in contemporary academic philosophy in the English speaking world and in Scandinavia, as “primarily concerned with the analysis of meaning” (Searle 1996, 2), as originating with Frege, Wittgenstein, Russell, and Moore, and as perpetuated by the logical positivists of the Vienna Circle and by the Oxford ordinary language movement. Similarly, Louis Pojman has recently claimed that “Analytic Philosophy is centered on language and logic, analyzing the meanings of words and sentences even as it analyzes arguments and builds comparatively modest epistemological and metaphysical theories” (Pojman 2001, 1). This, he says, constitutes a “simplistic but meaningful” characterization of analytic philosophy. Now these are exactly the sorts of definitions that, in Section 1, I claimed had been undermined by new wave scholarship. Indeed, this is why the perplexing variety of new definitions exists—because it has been shown that definitions involving traditional defining doctrines, in line with the received view, are not accurate. In the wake of this discovery, some have chosen to abandon the strategy of defining analytic philosophy according to doctrines, thereby implicitly rejecting the view that analytic philosophy is a philosophical school. This is a departure from the received view, but it is a subtle one. Prima facie, it allows us to preserve much of the received view, especially as concerns the extension of “analytic philosophy”. For instance, we saw earlier how defining analytic philosophy in non-doctrinal terms allowed Soames to focus on just the figures and thought-trends picked out as central and canonical on the received view, despite the fact that his canonical figures held no common views. Others, though, have departed from the received view in more conspicuous ways, ways that force the extensional scope of “analytic philosophy” well beyond its canonical domain. Definitions formed along these lines I call revisionist. Perhaps the most striking case of revisionism comes from Ray Monk. Taking a vague conception of analysis as the defining feature of analytic philosophy, Monk suggests that we carve up the philosophical world in such a way that Frege, Russell, Meinong and Husserl count as analytic philosophers while Wittgenstein does not (Monk 1996).7 To count Meinong and Husserl among the analysts while excluding Wittgenstein is unquestionably contrary to tradition—in fact, it is hard to imagine a definition more at odds with the canon derived from the received view of analytic philosophy. Another respect in which Monk’s view conflicts with the received view is that, since plenty of earlier philosophers used analysis in Monk’s sense, it detaches analytic philosophy from its customary turn-of-the-twentieth-century origin. Some revisionists acknowledge and accept this consequence. L. J. Cohen, for example, has argued that the analytic philosophers are united in that the problems they are interested in “are all, in one way or another, normative problems about reasons and reasoning, …” (Cohen 1986, 10 f.). But certainly interest in normative problems about reasons and reasoning is not unique to those who are commonly taken to be analytic philosophers. Cohen himself admits that, on his definition, analytic philosophy turns out to be “…a strand in the total history of western philosophy from Socrates onwards rather than just a modern movement” (Cohen 1986, 49). Similarly, Dagfinn Follesdal has defined analytic philosophy as philosophy with a strong commitment to argument and justification (as opposed to the kind of philosophy done by, e.g., Heidegger and Derrida, which relies mainly on rhetoric rather than clear argument), admitting that this makes Aristotle, Descartes, and perhaps even Thomas Aquinas count as analytic philosophers (Follesdal 1997).8 It seems to me that revisionism is misguided, for two reasons. First, it is self-undermining. The authors who end up proposing revisionist definitions do so only after using the received view to provide them with an initial orientation toward their subject matter. Starting off from the received view, they look for the defining doctrines of analytic philosophy. Finding none, they revise their conception of analytic philosophy in ways that utterly obliterate the received view, but then they carry on as if they had simply refined it. However, this involves what Putnam calls “excessive charity” in interpretation, similar to what would be required (borrowing Putnam’s example) to regard the concept of oxygen as a mere revision of the concept of phlogiston, rather than as a total replacement. Thus, on analogy with the phlogiston case, the original definiendum, analytic philosophy on the received view, doesn’t exist any more than phlogiston does. And, thus, it is clearly a mistake to carry on as if one has offered an improved definition of analytic philosophy, as the revisionists do. Moreover, this calls into question the whole line of thought involved in revisionism, since the first step taken in that line is guided by a conception that is not merely flawed, but false to the very core. Perhaps a more worrisome problem with revisionism is this practical one: by shifting the traditional boundaries of analytic philosophy both extensionally (in terms of who gets included or excluded) and temporally (in terms of when the school originated), it draws our attention away from the locus of the phenomena that explain analytic philosophy’s meteoric rise to power and prominence during the twentieth century—and this, I think, is what most needs to be explained by work in the history of analytic philosophy. As I have argued elsewhere (Preston 2005), and as can be gleaned from the proto-historical citations given earlier, this involved the widespread impression, itself originating and flourishing in the early-to-mid twentieth century, that there had been a philosophical revolution, complete with the emergence of a new, united philosophical regime. Assuming that analytic philosophy’s phenomenal social success in the twentieth century was to an appreciable extent due to the impression (which, from the standpoint of current scholarship, must be seen as a misimpression) that it was a united, revolutionary force armed with a powerful philosophical method (namely, the analysis of language), the problem with the revisionist strategy becomes clear: by detaching analytic philosophy from its turn-of-the-twentieth-century origins, it deprives us of any reasonable explanation for, first, analytic philosophy’s meteoric rise to power in the twentieth century, and, second, the fact that, even if there never was any real philosophical unity in analytic philosophy, it was for a long time thought that there was, and that it consisted in a metaphilosophical view according to which the nature of the philosophical enterprise was linguistic. The foregoing would seem to recommend the rejection of revisionist definitions. In doing so, however, it may seem that we have exhausted all our definitional options, with each type turning out to be a dead-end. Indeed those working in the history of analytic philosophy are presently confronted with a surprising quandary (though few, I think, have realized it, allowing themselves to escape from it too easily and before it fully emerges). The quandary can be expressed as follows. Contemporary historians begin their work with the following two assumptions:
(1) Analytic philosophy is a philosophical school. Both these assumptions are grounded in the received view of analytic philosophy as I have described it, and as represented by the proto-historians. However, recent scholarship has led many to the observation that: (3) There is no set of views accepted by all and only those figures ordinarily taken to be analytic philosophers (i.e., on the received view). Clearly, these propositions form an inconsistent triad, and one of them must be rejected. However, given that (3) is well supported, we cannot reject it; thus, the inconsistent triad reduces to a dilemma between (1) and (2). Rejecting (1) is the mark of a non-doctrinal definition; rejecting (2) is the mark of revisionism. But now, if (as per the arguments presented earlier) both these options are to be rejected, and if (3) demands the rejection of traditional definitions, what is left? What is left is the approach I call illusionist. On the illusionist view, we accept that the received view does not correspond and never has corresponded to anything in reality. Consequently, insofar as it has ever seemed to anyone that it did, that “seeming” was an illusion. More completely, the illusionist takes current work in the history of analytic philosophy to indicate that the received view was simply a guise that enabled a non-doctrinal and so non-philosophical group of some sort to come to dominate academic philosophy in various geographic regions by masquerading as a philosophical school.9 In this respect, the illusionist view can be characterized as rejecting (1): if analytic philosophy as ordinarily conceived is an illusion, then it is not a philosophical school, and (1) is false. And yet, the illusionist rejection of (1) does not qualify it as a variety of non-doctrinal definition; for the illusionist does not pretend, as those who offer non-doctrinal definitions do, that the lack of defining doctrines doesn’t matter to analytic philosophy’s nature as a philosophical school, and that the group represented by the received view can be recast as something lacking philosophical unity without destroying its philosophical nature and legitimacy. Instead, recognizing the centrality of the received view to the actual, historical developments associated with the name “analytic philosophy” (and vice-versa), illusionists allow it to exercise total control over the definition of analytic philosophy: for the illusionist, analytic philosophy is exactly what the received view says it is. In this respect, the illusionist view endorses a traditional definition. However, while other traditional definitions conflict with (3), the illusionist is saved from this precisely by treating analytic philosophy as an illusion. Thus, the illusionist is a traditionalist concerning what analytic philosophy is supposed to be, but differs from other traditionalists concerning whether analytic philosophy exists at all. This gives rise to a subordinate division in the category of traditional definitions: illusionist and benighted. Both adhere to the received view, but they differ in how they make use of it. Illusionism makes an enlightened use of the received view; it is analogous to the use that Wittgenstein made, and wanted his readers to make, of the propositions of the Tractatus: it is a ladder to be used to ascend to a higher plane of understanding, upon which one is freed from a kind of delusion concerning the meaningfulness of what one was doing previously—in this case, trying to unearth the true nature of analytic philosophy. Editing Tractatus 6.54 to fit our topic, we might say that The received view elucidates the true nature of analytic philosophy in this way: he who understands analytic philosophy according to the received view finally recognizes analytic philosophy as illusory, when he has climbed out through the received view, on it, over it. (He must so to speak throw away the ladder, after he has climbed up on it.) … He must surmount the received view; then he sees analytic philosophy rightly.Those who attain this elevated vantage point enjoy an illumination that frees them from the spell of the received view. They are no longer convinced that they are dealing with a type of philosophy. Consequently they feel no need to salvage something of the received view by searching, as the revisionists do, for some set of views to pull its raveling threads together into some really-existing philosophical school. Thus, illusionism is enlightened traditionalism. The alternative, unenlightened or benighted traditionalism, would be traditionalism which has not taken sufficient stock of the current state of research on analytic philosophy and its disturbing findings, and which, failing to climb out through, on, and over the received view, continues to operate under it.
* This essay is based on a talk entitled “Current Work on the History and Nature of Analytic Philosophy”, given at a joint session of the Bertrand Russell Society and the History of Early Analytic Philosophy Society, APA-Central Division Conference, Chicago, April 28, 2005. 1 Some works may occur out of sequence and others may not fit precisely into a category but be transitional stages with characteristics of several stages. These are exceptions, however; the general trend of development is well represented by this three stage schema. 2 One can also include in this category Nagel 1936a-b, Pap 1949, scattered reflections made in prefaces or introductions to works not otherwise historiographical (see, for example, the introduction to each volume of Ryle 1971), and critical observations made by opponents of analytic philosophy (such as Collingwood 1933, 1939, 1940, Gellner 1959, and Mundle 1970). Though not intended as history, all of these give us snapshots of analytic philosophy at various stages of development, and thus provide data for history in the same way that other proto-historical works do. 3 Cf. Hacker 1998, 4-14; Monk 1997; Hylton 1996, 1998; Beaney 2003. 4 For example, Ray Monk has argued that Russell’s “linguistic turn” in logic and mathematics (around 1912, at the hands of Wittgenstein) did not lead him to adopt a linguistic metaphilosophy; rather “the lesson he took from this is not that philosophers should now seek to analyze sentences, but rather that logic did not have, after all, the philosophical significance he had earlier attached to it” (Monk 1996, 56). 5 See Moore’s rejection of the linguistic interpretation of his own work in Moore 1942, 660 ff. It is noteworthy that this comes several years before Pap’s 1949 claims about the unity of analytic philosophy. In light of Moore’s protestations, it seems that it should have been impossible to say that Moorean analysis was the analysis of language; and yet, this is exactly what was asserted repeatedly by Pap and other proto-historians of analytic philosophy. 6 Both Soames’ indices and his treatment of canonical early figures like Russell and Wittgenstein suggest that he is largely unaware of new wave history (cf. Kremer 2005). Still, “ideas are in the air”, and Soames does exhibit a much greater degree of caution than Dummett about offering defining doctrines for analytic philosophy. 7 Monk is one of the few new wavers who, like Strawson in the protohistorical period, rejects the notion that differences over the nature of the objects of analysis should make a difference to how we group philosophers together as “analysts”. 8 It will be noted that Monk’s definition can be taken as methodological, while Cohen’s is topical and Follesdal’s stylistic. Given my views about non-doctrinal definitions, these would be non-starters. However, to allowsuch definitions as much mileage as possible, it is my policy to treat topical, methodological, stylistic, and (insofar as is practicable) other types of non-doctrinal definitions as containing implicit doctrinal definitions, and to take their doctrinal import as supplying their real significance. 9 Just what sort of group was behind the guise, and just how it carried off its masquerade, cannot be dealt with here. Preliminary answers are given in Preston 2005. REFERENCES
Ayer, A. J., et al. 1963: The Revolution in London: Macmillan.
Philosophy Department |