1 Introduction

Katzav and Vaesen have argued (Katzav 2018; Katzav & Vaesen, 2022) that part of the explanation for the emergence of analytic philosophy into dominance in the Anglophone world during the second half of the twentieth century was the use of institutional power by key analytic philosophers to marginalise non-analytic, and especially speculative, philosophy. This involved taking control of key journals, excluding non-analytic philosophy from their pages and founding analytic only journals. According to Katzav and Vaesen, the process of marginalisation started in the UK with the takeover of Mind in about 1925 and the founding of the analytic philosophy only journal Analysis in 1933. In the USA, the process of marginalisation started with the takeover of The Philosophical Review (PR) in about 1948, involved the takeover of The Journal of Philosophy towards the end of the 1950s and continued at least until the takeover of Philosophy and Phenomenological Research in the late 1970s. The pressure arising from control of journals appears to have peaked towards the end of the 1950s. In addition, the process of marginalisation was facilitated by control by logical empiricists of at least one source of research funding, namely the National Science Foundation’s History and Philosophy of Science funding program, and may have been facilitated by control of key philosophy departments, including, from the end of the 1940s, of the Sage School of Philosophy at Cornell and, from the 1950s, of the departments at Harvard and UCLA. The marginalisation of speculative philosophy continued while speculative philosophy was a substantial force, first in the UK in the 1920s and, later, in the USA in the 1950s. Furthermore, the marginalised material primarily included speculative philosophy but also included some forms of critical philosophy, such as the critical wing of phenomenology. What was preferred was analytic philosophy, conceived of as a form of critical philosophy. Indeed, Katzav and Vaesen propose that, in view of this institutionalised preference, analytic philosophy during the period 1925–1969 can be understood to be, at the institutional level, a sectarian form of critical philosophy.

Speculative philosophy (Katzav, 2018; Katzav & Vaesen, 2022) tends to go beyond established opinion in order to make claims that are to some extent independent of it, often offering a vision of reality as a whole in doing so. Critical philosophy tends to avoid going beyond established opinion, at least with respect to some parts of established opinion. Schools that are taken to be speculative include, among others, absolute idealism, personalism, American naturalism (including, e.g., classical pragmatism and Woodwardian naturalism), process philosophy and some forms of phenomenology. Schools that are taken to be critical include, among others, new realism, early analytic philosophy, mid-twentieth century linguistic philosophy, Quinean naturalism and some forms of phenomenology.

The evidence Katzav and Vaesen collect, however, primarily comprises journals coming under the control of analytic editors and then quickly shifting solely to publishing analytic philosophy. The evidence does not include evidence specifically for the thesis that, when departments came to be dominated by analytic philosophy, control was used to determine who got which jobs. And yet, if Katzav and Vaesen are correct, one would expect to see the pattern of sudden exclusion found in journals to be accompanied by a pattern of exclusion in hiring practices at influential philosophy departments in the USA. More specifically, if their theses about analytic philosophy are correct, one would expect the following theses to be correct:

  1. (1)

    After about 1948, there would be a relatively sharp, sustained increase in the ratio of hires of analytic philosophers by influential philosophy departments in the USA to the number of total hires of philosophers in those departments.

  2. (2)

    During the 1950s and 1960s, there would be a continued increase in the ratio of hires of analytic philosophers by influential philosophy departments in the USA to the number of total hires of philosophers in those departments.

  3. (3)

    Key departments in the USA would abruptly start more or less exclusively to hire analytic philosophers in the 1950s and 1960s.

  4. (4)

    Many students completing PhDs in the USA in the 1950s would still be writing on topics that were not associated with analytic philosophy at the time.

This paper aims to test predictions (1)–(4) against available data about job hires at influential philosophy departments in the USA during the twentieth century and against data about PhD topics in the USA during the period 1956–1965.

2 Methods

Jonathan Strassfeld (2020) compiled a database of hires by eleven prestigious mid-twentieth century philosophy departments, including those at Berkeley, Columbia, Chicago, Cornell, Harvard, Michigan, Pennsylvania, Princeton, Stanford, UCLA and Yale.Footnote 1 The database includes hires from prior to 1930 but does not include the years of hires for this period. The database does include hires by year and department for the period 1930–1979 (excluding Yale hires for the period 1975–1979). Furthermore, the database includes a classification of hires according to four categories: analytic, non-analytic, analytic/non-analytic and historical. I took the data for the categories analytic and non-analytic during the period 1930–1979 and aggregated it into numbers of hires per 5-year period (see Table 1). The result was used to test hypotheses (1), (2) and (3).

Table 1 Numbers of hires by department and year at influential USA philosophy departments. Strassfeld’s classification
Full size table

Strassfeld’s criterion for classifying philosophers was the extent to which they engaged in the discourses of analytic and non-analytic philosophy (2020, p. 857). For this reason, Strassfeld also includes a mixed, analytic/non-analytic category for those engaged in both discourses. By contrast, Katzav and Vaesen claim that the criterion determining which work was marginalised during the period under consideration is whether the work was in the analytic tradition, conceived of as a form of critical philosophy. Thus, this is the criterion relevant to testing their claims about marginalisation. Moreover, while a classification according to their criterion will include cases that are hard to classify, it will not include cases that count as analytic and non-analytic. Speculative philosophers generally recognised the importance of critical philosophy as part of philosophy. So, according to Katzav and Vaesen, anyone engaged in speculative philosophy counts as speculative, even if they also engage in critical philosophy, including analytic philosophy’s discourses.

In order to determine whether it is nevertheless appropriate to test Katzav and Vaesen’s theses using Strassfeld’s classification, I examined how his classification would be affected if revised in light of their criterion. I applied their criterion to the work of those on his list who were hired during the period 1930–1960. More specifically, I considered whether the body of work, including the explicit methodological commitments, of individual philosophers, would at the time be classified as the work of a speculative philosopher. This seems appropriate since my concern is with the causes of hires at the time. I then compared the trends in hires in the resulting classification with the trends in the one Strassfeld provided. My revised list also included revised classifications of the hires from 1930 to 1960 that Strassfeld classified as analytic/non-analytic, thus allowing me to consider the effect of my exclusion of this category from the data I used to test (1), (2) and (3). The process of revising the pre-1960s data, and selective examination of the classification of later hires, also allowed me to evaluate whether there was a need for a reclassification of post-1960s hires.Footnote 2

Let me offer some examples of reclassifications in order to better explain my reclassification procedure. Strassfeld classifies Andrew P. Ushenko as an analytic philosopher. This might seem plausible if we focus on which discourses he was involved in. Ushenko was a logician and a philosopher of physics, which are areas of research often associated with analytic philosophy, and he sometimes engaged with well-known figures in the analytic tradition. However, the philosophy of science, including the philosophy of physics, was a key area of research within speculative philosophy from at least the start of the twentieth century. The speculative philosophy of science aimed, among other things, to develop visions of reality by learning from, while also criticising and going beyond, science. Ushenko’s work aimed, and was recognised as aiming, at realising this goal. Moreover, his engagement with analytic philosophy, e.g. with logical positivism, was critical and aimed to defend speculative philosophy (Reck, 1958; Katzav & Vaesen, 2022). He is thus classified as a speculative philosopher by me. Morris R. Cohen, who is classified as analytic/non-analytic by Strassfeld, is also reclassified as speculative by me. Cohen was particularly influential in developing and spreading speculative philosophy of science (Katzav & Vaesen, 2022). Walter T. Stace, on the other hand, is classified as non-analytic by Strassfeld but as analytic by me. While many of the views and authors Stace discusses are outside of the analytic tradition, he discusses them with the aim of establishing the boundaries of established opinion, e.g. whether mysticism should be part of it. He is also concerned with unpacking or analysing parts of established opinion rather than with going beyond them (Stace, 1960; Katzav & Vaesen, 2017, p. 784).

The Review of Metaphysics (RM) published data about PhD completions in philosophy at key universities in the USA and Canada during the period 1956–1965.Footnote 3 The data includes information about the number of PhDs awarded by each department, the names of the students being awarded PhDs, the titles of their PhDs and the names of PhD supervisors.Footnote 4 The RM data includes data about PhDs from all of the already mentioned key departments, except for Stanford. The RM data was used to provide a qualitative test of hypothesis (4) about non-analytic PhDs. Reading off the philosophical orientation of a thesis from a title is a risky matter. For example, in the 1950s, there were still a substantial number of pragmatists and speculative philosophers who were working in the philosophy of science (Katzav & Vaesen, 2022), so that a title such as ‘An examination of some aspects of natural science’ could just as well be in the non-analytic tradition despite the common association of philosophy of science with analytic philosophy. Nevertheless, some titles are more likely to feature in non-analytic theses, and titles were combined with supervisor orientation to suggest thesis orientation.

3 Results

The data from Strassfeld shows a sharp increase in the ratio of analytic to non-analytic hires at about 1950, one that comes after a period of relative stability of this ratio and reflects a very substantial change in hire trends (Fig. 1). Prior to 1950, hires of analytic philosophers fluctuate between about 30 and 40% of the total number of hires. Immediately after 1950, analytic hires comprise about 70% of all hires. The ratio continues to grow in the 1950s and, by the 1960s, analytic hires comprise between about 80 and 90% of all hires. This post-1950 change in growth is also substantial.

Fig. 1
figure 1

Analytic to non-analytic hires (Strassfeld’s classification)

Full size image

There are some discrepancies between my classification, based on Katzav and Vaesen’s work, and Strassfeld’s classification. In particular, the jump that occurs immediately after 1950 is not quite as steep, on my classification.Footnote 5 Nevertheless, the jump is still steep, and the differences between the classifications do not alter the trends in a way that is relevant to the qualitative judgments made here (compare Fig. 1 and Fig. 2). Furthermore, at least by the 1940s, the number of cases in which reclassification occurred, including reclassification of mixed analytic/non-analytic cases, was sufficiently low relative to the total number of hires so as to make no real difference to the overall trend of hires (compare the final rows of Table 1 and Table 2). The proportion requiring reclassification does not appear to increase in later years. Finally, the cases of reclassification are roughly evenly distributed between cases being reclassified as non-analytic and cases being reclassified as analytic. It thus seems that alterations to Strassfeld’s data after 1960 are not going to alter the present paper’s conclusions.

Fig. 2
figure 2

Analytic to non-analytic hires (this paper’s reclassification of Strassfeld’s data, 1930–1959)

Full size image
Table 2 Numbers of hires by department and year at influential USA philosophy departments. Classification based on this paper’s application of Katzav and Vaesen’s criterion
Full size table

The shift to hiring more or less only analytic philosophers is suggested by the data for all listed departments, with the exception of Yale and perhaps, to a limited extent, Berkeley (see Tables 1 and 2). In the case of Cornell, the shift appears to occur immediately after the hiring of the speculative philosopher Arthur E. Murphy in 1945. At Columbia, the shift appears to occur in about 1960. Similar shifts appear to be found at Harvard (from about 1950) and Chicago, UCLA and Princeton (all from about 1955). Pennsylvania only hires analytic philosophers from 1960 onwards but does not appear to have had any hires in the 1950s. Stanford ceases to hire non-analytic philosophers from at least 1960. Michigan’s focus is solely on analytic philosophy since about 1930.Footnote 6

Contrary to what one might expect, but in accord with Katzav and Vaesen, the RM data on PhD completions does not suggest that these are dominated by analytic topics and supervisors. This is very clear when looking at the totality of departments but is so even if we limit our attention to the prominent departments that also appear in Strassfeld’s list. By far, the departments producing the most PhDs during the period 1956–1960 from among those on the list are Columbia and Yale. They produce 107 out of the 199 PhDs listed for this period (see Table 3). During these years, many of the theses from both departments appear to be non-analytic when judged on the basis of title and supervisor (see Table 4 for illustrative examples). Chicago too appeared to be producing a substantial number of non-analytic theses until 1960 (see Table 4). Indeed, the total number of PhDs written at departments which are on Strassfeld’s list and appear to be predominantly producing, or only producing, analytically oriented theses in the 1950s is relatively small. These departments were Harvard, Michigan, Princeton, UCLA and Berkeley and appear to have graduated only 68 PhDs between them during the period 1956–1960, with 32 being from Harvard (Table 3).Footnote 7 Columbia, Yale and Chicago graduated 124 PhDs during the same period.

Table 3 Numbers of PhDs in philosophy by university and year (1956–1965)
Full size table
Table 4 Illustrative PhD topics from Chicago, Columbia, Harvard and Yale
Full size table

4 Discussion

In accordance with (1), the data examined here indicate that there was a sharp increase in the ratio of hires of analytic philosophers to non-analytic philosophers by influential departments from about 1950. In accordance with (2), this ratio continued to increase during the 1950s and 1960s. (1) and (2), recall, are what we would expect if the takeover by analytic philosophers of the Sage School of Philosophy and of PR and subsequent takeovers of other journals and academic institutions in the 1950s play a substantial role in explaining the emergence into dominance of analytic philosophy.

In accordance with (3), the data suggest that influential departments start more or less exclusively to hire analytic philosophers in the 1950s and 1960s. While the hires at individual departments are low, the similarity of patterns across departments supports the existence of a move to a more or less exclusive preference. Again, this is to be expected if institutional control is a key driver of the growth of analytic philosophy during these decades. Finally, in accordance with (4), many students completing PhDs in the 1950s were still working outside of what was then analytic philosophy. There is thus additional evidence, in line with the evidence from the contents of journals and books, that non-analytic philosophy was still a substantial force in American academia.

In fact, the data suggest stronger hypotheses than those put forward by Katzav and Vaesen. The takeover of Cornell by analytic philosophers gave analytic philosophy control over one of the key organs of American philosophy, namely PR (Katzav, 2018). But Cornell does not appear to have produced many PhD students in the 1950s, so it alone could not make possible a dominant analytic tradition. However, Harvard’s commitment to analytic philosophy, which seems to be in place from at least 1950, gave the analytic tradition in America a substantial source of analytically inclined philosophers.Footnote 8 Plausibly, widespread knowledge of the shifts at Cornell and Harvard encouraged further preference for analytic philosophy. By 1955–1959, Harvard’s PhD producing capacity was supplemented by those of Chicago and Princeton, which are then only hiring analytic philosophers. At this stage, non-analytic philosophy not only has limited venues for publication but also extremely limited influential destinations for those completing PhDs; the destinations are (throughout the 1950s and 1960s) Yale and (during the 1950s) Columbia and, to a limited extent, Berkeley. Moreover, Berkeley is predominantly analytic, while Yale and Columbia are open to analytic hires, thus further giving an advantage to analytic philosophy. It accordingly seems that control of journals, PhD production and hires meant that, by the middle of the 1950s, analytic philosophy was bound to come out on top, at least when it comes to the most influential American departments. When hire numbers really rocket in the 1960s, only analytic philosophers would have a chance at finding placements at all but one of the influential departments considered here.

It is important to emphasise that there may well be other, not necessarily incompatible with the one just given, explanations for the growth of analytic philosophy in America in the 1950s and later. Two of the available explanations for this emergence are relevant here, since they identify events shortly before and during the 1950s as causes of this growth. Some explain the 1950s growth by pointing to the arrival of the logical positivists in America, starting in the 1930s, to inherent affinities between analytic philosophy and American philosophy prior to 1950 and to the rapid expansion of Academia after 1950 (see, e.g., Soames, 2008). While there is plausibly something to this explanation, it does not explain the very rapid shift to an almost total preference for analytic philosophy. Indeed, this explanation exacerbates the puzzle about the success of analytic philosophy. The logical positivists were small in number while there was an existing, larger community of philosophers of science in America. Moreover, the American group was predominantly speculative and thus had in-principle objections to positivism’s anti-metaphysics stance (Katzav & Vaesen, 2022). It is unclear why the American tradition would simply give way to a small group with which there was fundamental disagreement about philosophical approach. Furthermore, it is unclear why the expansion of academia in the 1950s and beyond would, by itself, strongly favour analytic philosophy. Indeed, non-analytic philosophy did benefit from the growth in academia immediately after the war (see Tables 1 and 2).

John McCumber (2001) argues that the political pressures of McCarthyism favoured analytic philosophy over more politically engaged rivals, such as pragmatism. Given that McCarthyism starts to bite in 1949, McCumber’s hypothesis does lead us to expect a substantial increase in hires of analytic philosophy after this time. So, his explanation does explain some of the data provided here. That said, an explanation solely appealing to McCarthyism does not explain the data as well as one that appeals solely to institutional control. The changes at Cornell look like they precede 1949, something independently supported (Katzav & Vaesen, 2017) by the fact that PR starts excluding non-analytic philosophy from its pages in 1948. The changes at Harvard coincide with McCarthyism, suggesting that the groundwork for what happened there, groundwork such as the hiring of individuals committed to hiring only analytic philosophers, precedes McCarthyism. Similarly, McCarthyism’s influence was too short-lived to fully explain the continued preference for analytic philosophy in the second half of the 1950s and in the 1960s, so that, at a minimum, further factors are needed to explain why its effects persisted then. McCumber need not disagree about the limitations of appeals to McCarthyism; he does not claim it suffices to explain the dominance of analytic philosophy. Still, the data helps to indicate the extent to which factors other than just McCarthyism are needed to explain this dominance.

In summary, Strassfeld’s data and my revision of this data are fully explained by the hypothesis that analytic philosophy gained substantial ground from about 1950 and did so as a result of institutional control. The data also fit well with the proposal that, at the institutional level, analytic philosophy was a sectarian form of critical philosophy, at least during much of the twentieth century. To be sure, other explanations for the data cannot be excluded. An appeal to the affinity between parts of American philosophy with analytic philosophy can play some explanatory role here, but one substantially limited by the existence of disagreements about philosophical approach between analytic and speculative philosophers. McCarthyism too can play a substantial role in explaining the data, though they also suggest that such an explanation can only be partial.