“For the perfections of Language, not properly understood, have been one of the chief causes of the imperfections of our philosophy.” (John Horne Tooke ‘The Diversions of Purley’ p. 19)

The above quote, taken from ‘The Diversions of Purley[1]’, demonstrate John Horne Tooke’s views on language and its relation to philosophical thinking; and anticipate views which would become prominent in twentieth century analytic philosophy. In particular the above quote would put one in mind of the work of the later Wittgenstein. There is no evidence that Tooke’s work in anyway influenced Wittgenstein’s philosophy. However, we do know that Tooke’s work influenced the philosopher and logician Willard Quine and psychologist B.F. Skinner.

Skinner came across Tooke’s work in the early thirties when he was a junior fellow at Harvard:

“Henderson urged me to look at John Horne Tooke’s ‘Diversions of Purley’…the book was out of print but I advertised, and several booksellers sent me quotations. I brought two and gave one to Van Quine inscribed Verbum Sat.” ( Skinner “The Shaping of a Behaviourist” p. 158) (ibid p. 282)

In his autobiography ‘The Time of My Life’, Quine recalled Skinner giving him a copy of Tooke’s book:

“It was particularly in language theory, rather, that Fred opened doors for me. My linguistic interest had run to etymological detail; he put me onto Bloomfield and Jesperson and gave me an American first edition of John Horne Tooke.” (Quine: The Time of My Life p. 110)

In this blog-post I will discuss the influence that Horne Tooke’s book had on both Quine and Skinner, and what their respective reactions to Tooke, reveals about their different behaviourist philosophies.

Quine and John Horne Tooke

In 1946 Quine gave a series of lectures on the philosophy of David Hume. In the lecture series he obviously related Hume to the other empiricists and rationalists who were contemporaries and near contemporaries of Hume. During these lectures Quine discussed the work of John Horne Tooke, who Quine believed, had made an advance over the British Empiricists. When discussing the British Empiricists Quine noted that they were all wedded to the idea idea conception of epistemology. Tooke’s ‘Diversions’ was an attempt to move away from this idea centric epistemology. Tooke considered and critiqued the work of John Locke but didn’t discuss either David Hume or George Berkeley’s work.

Tooke argued you could translate Locke’s talk of ‘ideas’ with talk of ‘words’ and you would increase the clarity and correctness of Locke’s philosophy. Quine agreed with Tooke’s assessment of Locke’s philosophy, and thirty years later in his paper ‘Five Milestones of Empiricism’ argued that Tooke’s move away from idea centric philosophy to an emphasis on words was one of the key milestones in the development of empiricist philosophy.

Tooke’s philosophy reduced all discourse to two main categories; nouns and verbs. He argued that one could explain other linguistic phrases such as ‘prepositions’, ‘adjectives’ etc by analysing them. Upon analysis he claimed that such words contained a hidden complexity. Thus, for example, Tooke analysed the preposition ‘for’ interms of the underlying notion of ‘cause’. He did this by analysing an incredible amount of sentences containing the word ‘for’ and showing how the sentences could all be correctly analysed by treating ‘for’ as meaning ‘cause’. Thus Tooke was satisfied that he could analyse away the preposition ‘for’ and treat it as a verb. He used similar processes of analysis on a variety of other prepositions, and on adverbs, conjunctions etc. As well as analysing various different parts of written language to reveal its function; Tooke also tried to explain how these words evolved over time by examining their etymology.

Quine admired what he called Tooke’s the method of abbreviations (Quine ‘Lectures on the Philosophy of David Hume’ p. 62). In his ‘Divergence’ Tooke argued that when trying to understand speech we need to conceive of it as words which are necessary to communicate our thoughts and abbreviations which help with expressing these thoughts clearly. Tooke argued that there are two sorts of words necessary to the communicating of our thoughts; nouns and verbs. Everything else he conceived of as being abbreviations which when analysed closely will be shown to be either nouns or verbs. His analysis of the word ‘for’ above is a good example of his understanding his method of abbreviation. In his 1951 ‘Two Dogmas of Empiricism’ Quine related this method of abbreviations to the verification theory of meaning:

“Radical reductionism, in one form or another, well antedates the verification theory of meaning explicitly so called. Thus Locke and Hume held that every idea must either originate directly in sense experience or else be compounded of ideas thus originating; and taking a hint from Tooke we might rephrase this doctrine in semantical jargon by saying that a term, to be significant at all, must be either a name of a sense datum or a compound of such names or an abbreviation of such a compound. So stated the doctrine remains ambiguous as between sense data as sensory events and sense data as sensory qualities; and it remains vague as to the admissible ways of compounding. ( Quine ‘Two Dogmas of Empiricism’ p. 38)

John Locke believed that we begin with simple ideas derived from perception and combine them (somehow) to form complex ideas when thinking. Locke further argued that the words in our language got their meanings by referring to these ideas. What Quine admired about Tooke’s work was that he cut out the middle man so to speak. Tooke was emphasising the fact that our words got their meanings in by picking out things and events in the environment. On Tooke’s picture ‘ideas’ were a theoretically superfluous posit:

“Every purpose for which the composition of Ideas was imagined being more easily and naturally answered by the composition of Terms: whilst at the same time it does likewise clear up many difficulties in which the supposed composition of Ideas necessarily involves us.” (Hooke ‘Divergences’ p. 20)

Quine asks us to note that we shouldn’t read Tooke’s criticism of complex ideas as denying the importance of mental activity, nor should one think that the concept of complex definitions (abbreviations) don’t involve mental activity. Rather, Tooke was just pointing out that ideas as explanatory posits don’t do much work in clarifying how we connect stimulation with discourse (Quine ‘Lectures in the philosophy of Hume p. 63). In his 1977 paper ‘Facts of the Matter’ Quine made the point as follows:

“Let us therefore recognize that the whole idea idea, abstract and concrete, is a frail reed indeed. We must seek a firm footing rather in words. The point was urged by John Horne Tooke only shortly after Hume’s time, in 1786. Tooke held that Locke’s essay could be much improved by substituting the word ‘word’ everywhere for the word ‘idea’. What is thereby gained in firmness is attended by no appreciable loss in scope, since ideas without words would have come to little in any event. We think mostly in words, and we report our thoughts wholly in words. Let us then take one leaf from the old-time philosophy and another from John Horne Tooke. Philosophical inquiry should begin with the clear, yes; but with clear words. (‘Facts of the Matter’ p. 271)

In Quine’s ‘Five milestones of Empiricism’ (1978), he again, credits Tooke with emphasising the importance of words over ideas, arguing that this move was a key milestone in the development of empiricism. The other four milestones Quine discusses are Bentham’s emphasis of the sentence having semantic primacy in language over words, Duhem emphasis of the primacy of systems of belief over sentences, his dissolution of the analytic/synthetic distinction which he argues this leads to methodological monism, and his demonstration that there is no first philosophy.

I won’t here speak of his last four milestones of empiricism, given that the subject matter of the blog is John Horne Tooke, I will focus on Quine’s first milestone of empiricism. Quine noted the following:

“The first was the shift of attention from Ideas to words. This was the adoption of the policy, in epistemology, of talking about linguistic expressions where possible instead of ideas…I think of it as entering modern empiricism only in 1786, when…John Horne Tooke wrote as follows: “the greatest part of Mr. Locke’s essay, that is.  All which relates to what he calls the abstraction, complexity, generalization, relation etc., of ideas, does indeed merely concern language.” British empiricism was dedicated to the proposition that only sense makes sense. Ideas were acceptable only if based on sense impressions. But Tooke appreciated that the idea idea measures up poorly to empiricist standards. Translated into Tooke’s terms, then, the basic proposition of British Empiricism would seem to say that words make sense only insofar as they are definable in sensory terms” (“Five Milestones of Empiricism” p. 68)

Quine notes that this approach of Tooke’s leads instantly to problems. The grammatical particles which we use to organise our concepts don’t easily reduce to sensory experiences. As we saw above Tooke tried to avoid this problem by saying that sentences could be reduced to two functions ‘nouns’ and ‘verbs’; thus nouns refer to sensory experiences while verbs say things about these experiences. To work within this austere empiricist frame work Tooke had to explain away grammatical concepts such as ‘if, and, but, there’ etc interms of nouns and verbs. Tooke justified this approach by giving unpersuasive etymological definitions of these grammatical concepts.

Quine was pretty dismissive of Tooke’s attempts to explain the grammatical concepts interms of nouns and verbs. He argued cogently that Tooke didn’t realise that these concepts were syncategorematic; they couldn’t be defined in isolation but only in context.

Quine was largely correct in his argument that grammatical concepts are not definable in isolation. But he didn’t sufficiently appreciate the possibility that grammar may be an innate imposition on how we group words together. Quine was working in the logical positivist tradition which worked to reduce our theories to sensory experiences and logical constructions based on sensory experience. While he was correct that grammatical concepts cannot be defined in terms of sensory experience and are syncategorematic; he seems to entirely ignore the possibility that grammatical concepts be indefinable (by which he means they cannot be explained interms of sensory experience), because they are innate and are used in helping us interpret sensory experience. In short in his discussion of the five milestones of empiricism Quine was guilty of underplaying the role of non-empirical Kantian (or at least Chomskian type knowledge). I am not arguing that the Kantian/Chomskian alternative is the correct explanation of the grammatical particles. I am just noting that his empiricism is blinding him to an alternative explanation. And this blindness is particularly interesting to note given that Quine noted many times in his interactions with Chomsky that he had no difficulty with explanations which appealed to innateness.

However, it is not within the scope of this particular blog-post to discuss the evidence for innate syntax so I will not pursue the above criticism of Quine here. The key point to note is that while Quine agreed with Horne Tooke’s movement from explanations in terms of ideas to explanations in terms of terms; Quine didn’t agree with Tooke’s analysis of grammatical particles. In the next section I will explore how Skinner deals with Tooke’s analysis of grammar and Tooke’s criticisms of idea centric philosophy.

Skinner and John Horne Tooke

 “The French novel of the nineteenth century was possibly close to what I wanted, and I reread Stendhal and Balzac. I was caught up in a renewal of interest in George Eliot and tried rewriting parts of ‘Middlemarch’ and ‘Daniel Deronda’, replacing references to feelings with references to the actions from which feelings were inferred. It did not work. Mentalistic terms were like the “abbreviations” of John Horne Tooke; they were the products. Accurate reports of the same contingencies ran to much greater length.” (B.F. Skinner ‘A Matter of Consequences’ p. 245)

1. F. Skinner began working on whether language could be explained behaviouristically in the mid-nineteen thirties after a challenge set to him by the philosopher Alfred North Whitehead. For over fifty years after his discussion with Whitehead, Skinner at various times in his career worked on the nature of language. Throughout this fifty year period whenever he discussed language the name of John Horne Tooke came up. In his 1947 lectures on language (later called the Hefferline notes), Skinner briefly spoke approvingly of Horne Tooke’s work:

“John Horne Tooke, Englishman of the 18^th century, wasn’t liked and was popped into jail once or twice by the government. He had one trial which hinged on the interpretation of the word ‘that’. This got him going and he wrote a book…He was a good behaviourist although he didn’t know it.” (Skinner ‘The Hefferline Lectures’ p. 21)

Unfortunately despite speaking approvingly of Horne Tooke in ‘The Hefferline Lectures’ Skinner didn’t expand on what it was about Horne Tooke’s work that he found impressive. Forty years later when discussing the evolution of ‘Verbal Behaviour’, Skinner again mentioned Horne Tooke’s work:

“An early effort by John Horne Tooke in the ‘Diversions of Purley’ (1776) has not been fully appreciated. That Tooke was not always right as an etymologist was not as important as his efforts to explain how English speakers could have come to say such words as ‘if’, ‘but’, or ‘and’.” (‘The Evolution of Verbal Behaviour’ p. 120)

It is no coincidence that Skinner’s interest in Horne Tooke centred on his analysis of concepts such ‘if’, ‘but’, ‘that’, ‘and’ etc. Skinner was also impressed with Quine’s analysis of similar concepts in his ‘Elementary Logic’. In fact in ‘Verbal Behaviour’ where Skinner discussed Horne Tooke in most detail he notes that Horne Tooke’s analysis of language was similar to Quine’s analysis in ‘Elementary logic’ (Verbal Behaviour p.342).

It was in his ‘Verbal Behaviour’, that Skinner discussed John Horne Tooke in most detail. Interestingly Skinner’s discussion of the Tooke was along the same lines as Quines. Skinner, like Quine, discussed Horne Tooke’s criticism of Locke’s ‘Inquiry into Human Understanding’ for being better thought of as being concerned with words rather than ideas. Skinner even cites the same passages from Horne Tooke re-John Locke that Quine did. However, while Quine was careful to note that Tooke was primarily speaking about language over ideas because ideas were non-explanations, he also noted that we shouldn’t read Tooke as denying that mental activity underlay verbal behaviour. Skinner on the other hand read Horne Tooke as arguing that all thinking involved verbal behaviour; Skinner then goes on to argue that Tooke is incorrect in arguing thusly and points to things such as mental imagery, and spatio temporal reasoning as a refutation of Tooke’s purported views (‘Verbal Behaviour’ p.449).

Skinner and Quine’s different readings of Horne Tooke are understandable. Horne-Tooke is very articulate on what he sees as the problems with idea-centric philosophy. He also has skilled arguments for using terms and their analysis and historical development to make our philosophy more objective. But he says little (either positive or negative) about whether he thinks that there is cognitive apparatus underling the ability to use verbal behaviour. So there is scope for both Quine and Skinner to differ in their interpretations of Horne Tooke on this issue and little textual data to settle the matter conclusively.

Another area where Skinner and Quine discussed Horne Tooke was in relation to his treatment of grammar. Skinner was particularly interested in Horne Tooke in relation to what he called autoclitics. Before proceeding to discuss Skinner’s take on Horne Tooke re-autoclitics I will need to briefly discuss Skinner’s explication of the various different functional units that make up ‘Verbal Behaviour’. A Mand is a Verbal Operant in which the response is reinforced by a particular consequence; and hence is under the functional control of relevant conditions of deprivation or aversive stimulation (‘Verbal Behaviour’ p. 36). In other words the mand is a type of response that is under the control of and singled out by certain controlling variables. A paradigm of a mand is saying ‘water’ when thirsty and receiving water in return (being reinforced for saying ‘water’).

A tact is a verbal operant that is controlled by non-verbal stimulus. The child says ‘doll’ in the presence of a doll and is reinforced. Used as a mand the word ‘doll’ would result in the child being handed a doll. But as a tact the child says the word ‘doll’ in the presence of a doll and is reinforced by his peers (through praise, attention etc).

An echoic is Verbal Behaviour that is controlled by other Verbal Behaviour. Thus the child repeats the word ‘doll’ upon hearing the word ‘doll’ spoken. Intraverbal Behaviour is behaviour where Verbal Behaviour is controlled by other Verbal Behaviour; but where the there isn’t a formal correspondence between the stimulus and response product (Verbal Behaviour p. 71). An example of echoic behaviour would be one person saying ‘the wheels on the bus’ and the other person saying ‘the wheels on the bus’. Whereas, an example of an Intraverbal behaviour; would be one person saying, ‘The wheels on the bus’, and the other person saying ‘go round and round’. Skinner uses Intraverbal behaviour to explain analytic truths. ‘Thus 2 plus 2’ ‘equals 4’ would be explained as an Intraverbal where 4 is under the control of 2 plus 2.

The autoclitic is a form of Verbal Behaviour that modifies other verbal operants such as the mand, the tact etc. Skinner notes that there are different types of autoclitics. One type is the descriptive autoclitic which says something about the particular verbal operant that is used; so if you take the word ‘heads’ this can be modified by a descriptive autoclitic as follows (I said (heads)), (I will say (heads)) etc. There are many different sub-types of descriptive autoclitics such an autoclitics with indicate my strength of belief in a verbal operant I have emitted; thus I could modify the tact ‘the cat is black’ with the autoclitic of weakness (I hesitate to say (the cat is black).

As well as descriptive autoclitics Skinner also discusses qualifying autoclitics, quantifying autoclitics and manipulative autoclitics. It was in relation to autoclitics that Skinner discussed John Horne Tooke’s work.

As we saw above Horne Tooke was concerned with explicating language in terms of nouns and verbs. Tooke believed that he could explain away the other aspects of language by analysing them as being abbreviations which ultimately were nouns or verbs. Horne Tooke’s method was drawing out the terms meanings through analysis, and explain how the terms had the form they did by tracing their etymology. Thus when analysing the preposition ‘through’ Horne Tooke analyses it as deriving from the nouns ‘door’/ ‘gate’/ ‘passage’; his justification is dual. He shows how he can analyse common uses of ‘through’ interms of ‘door’/’gate’/’passage’ and he traces the etymology of the term ‘through’ to justify his analysis (‘Divergence’ pp.180-183).

Tooke’s analysis is interesting and puts one in mind of the work of Lakoff and Johnson who analyse our language as deriving from embodied experiences to more abstract realms. Thus a common physical object  such as  a door or a gate that we have an embodied relation to are used in more abstract senses to think about more complex objects. It is not within the remit of this blog-post to evaluate the truth of Tooke’s analysis rather I just want to trace what Skinner and Quine made of Tooke’s views.

Skinner admired Tooke’s analysis of language, however he didn’t agree with Tooke’s contention that all language could be reduced to verbs and nouns. As we saw above Skinner didn’t analyse language interms of traditional grammatical categories; rather he argued that the key to understanding language was to analyse it in terms of various type of behavioural functions (mands, tacts etc). Skinner noted that Tooke’s analysis was hindered by the fact that Tooke had no real understanding of the fact that some words were used to deal with other parts of language. According to Skinner, Tooke’s abbreviations were just words which were used to manipulate nouns and verbs, and not grasping this fact held back Tooke’s analysis of language (Verbal Behaviour p. 341):

“What Tooke lacked was a conception of behaviour as such. He was still under the influence of British empiricism and, in spite of an heroic declaration of independence…Struggling against an enormous weight of tradition, Tooke is talking about verbal behaviour. He has “disabbreviated” the puzzling terms which cannot be accounted for as object words or by appeal to images-terms which we would classify here as autoclitics- and  has found that they are verbs. This leads him to an important generalization which we could paraphrase in this way: some verbal responses are evoked by external state of affairs. These Tooke wants to call nouns. Other responses are communication itself. They affect the listener and have no function aside from that effect. Tooke wants the listener to have no function aside from that effect. Tooke wants to call them verbs. Writing more than a hundred and fifty years ago, he had no alternative, but a fresh formulation is possible today.” ( Skinner ‘Verbal Behaviour’ p. 343)

Skinner was impressed with Tooke’s recognition that language had a dual function; referring to objects in the external world; and communicating about these objects via verbs. However, Skinner noted that as a thinker of his time Tooke didn’t have a sufficient grasp of the various different functions of language and the social reinforcement controlling these behavioural functions.

Both Skinner and Quine were impressed with Tooke’s move away from Locke’s idea idea epistemology. Though they interpreted Tooke’s move in different ways; Quine seemed to believe that Tooke’s views were compatible with a mild form of cognitivism though not of the sort that would vindicate folk-psychology. While Skinner read Tooke as overplaying the linguistic nature of thinking. When it came to grammar both Quine and Skinner, while impressed with Tooke’s work, had some reservations. Quine argued that Tooke didn’t appreciate the contextual nature of grammar and erroneously tried to reduce them to sensory impressions and judgements about these impressions. Skinner on the other hand disagreed with Tooke’s grammar because he didn’t think that Tooke sufficiently appreciated the various different functions of language; nor the reinforcing contingencies that shaped these functions.

Skinner and Quine’s different criticisms of Tooke aren’t necessarily incompatible but they do illustrate their divergent interests. Quine the great critic the idea that our epistemic contact with the world can be purely cashed out in sensory terms; railing against Tooke’s attempt to explain our linguistic capacities in terms of sensory experiences. And Skinner attempting to explain language behaviourally and functionally, admiring Tooke’s attempts to step out of the Cartesian Tradition he was trained in, but lacking an account of behaviour powerful to complete the job.

[1] Henceforth I will refer to ‘The Diversions of Purley’ as ‘Diversions’.