8. J.R. Firth¹ 

8.1 Although John Rupert Firth helped to found linguistics in Great Britain, he published no major theoretical book. He wrote two short popular books early in his career, intended to arouse general interest in language study -- Speech (1930) and Tongues of Men (1937) -- but otherwise only assorted occasional papers eventually collected in two volumes, one posthumous. According to Frank Palmer, who edited the second, Firth was rumoured to be ‘preparing a book entitled Principles of Linguistics’, but ‘among the papers he left at his death was not one sheet’ of it (P2 2). So I will essay to reconstruct Firth's theory and method from the four extant volumes, whose materials were designed for a variety of general or specific purposes and audiences. Despite vacillations and contradictions,² Firth's position remained fairly consistent during the three decades summarized here. The popular books, according to Peter Strevens (editor of their 1964 re-issue), already contained ‘the seeds of a great many concepts’ used in Firth's ‘subsequent academic works on linguistics’ (TMS viii).

8.2 In Palmer's estimate, Firth ‘alone pioneered’ ‘linguistics’ ‘in Britain’, developing ‘his own original brand’ and many ‘exciting new ideas’ (P2 1). Firth held ‘the first chair in general linguistics’ in England, which ‘was established in the University of London in 1944 at the School of Oriental and African Studies’ (P1 v, P2 96). He remained, Strevens says, largely ‘unknown to linguistics in the American tradition’ (TMS vii), aside from Pike (cf. Kachru 1980);³ in the U.S., ‘British ideas’ were deemed ‘a variant of Bloomfieldian linguistics’ or even ‘a deviant consequence of having misunderstood American linguistics’ (TMS ix; cf. P2 2). Firth did espouse some ideas also encountered in Bloomfield,⁴ but mainly due to ‘the intellectual climate’ and ‘the context of science’ in the Anglo-Saxon world at the time (cf. P1 169). Sometimes Firth showed solidarity with American linguists, but other times depicted their work as narrow or misguided and fundamentally different from his own.⁵

8.3 Unlike our other theorists (even pious Pike), Firth salutes ‘the importance of religion’ ‘in the history of Western linguistics’ alongside ‘science’, especially ‘the Christian missions like the Roman Catholic “Propaganda Fide”‘ and the ‘Summer Institute of Linguistics’ that ‘trains missionaries’ (TMS 11, 136, 138, 55f, 59, 107, P1 164, P2 162) (cf. 5.2). At ‘the Third International Congress of Linguists’ in 1933, the Pope ‘said that the whole redemption was the work of the “Word”‘, and Firth alludes to legends and holy books portraying language as the invention of a god (TMS 13, 3-6, 15f). But he recognizes ‘religious and linguistic expansion’ as a ‘supplement’ to ‘more material interests’ (TMS 59). After all, ‘world languages’ are ‘built on blood, money, sinews, and suffering in the pursuit of power’ and made by ‘men of action’: ‘statesmen, soldiers, sailors’, not just ‘missionaries’ (TMS 71). Firth is thus sceptical about ‘universal languages’ invented by ‘amateur grammarians’ and ‘linguists trying to undermine Babel’, the ‘most successful’ being Esperanto (TMS 70, 11, 49, 66, 68) (cf. 8.6). He wants to reserve the role for English (8.12).

8.4 Being an Orientalist, Firth also salutes the dawn of language study in India and ‘the discovery of Sanskrit by the West’ (TMS 147, P1 111, P2 114, 1168; cf. 2.5; 4.4, 4³; 8.74).⁶ Admittedly, ‘the ancient Hindu grammarians, and later the Arabs, were not interested in vernacular’ but in ‘preserving the purity of sacred languages from vulgar mutilation and defilement’ (TMS 147) (cf. 4.40). Much ‘word-craft’ has similarly been ‘transmitted by privileged elites: elders, priests, clerks, sheiks, mandarins, bureaucrats’; some ‘men of “letters”‘ even ‘became rulers’ (TMS 48, 146). Their descendants are ‘all those who believe in arbitrary linguistic standards’ and ‘purity’, and adopt a ‘static, mainly prescriptive or normative’ ‘point of view’ from which ‘linguistic evolution’ gets ‘evaluated’ as ‘decay’, ‘degeneration’, or ‘corruption’ (TMS 147f) (cf. 2.49).

8.5 Still, Firth thinks ‘the great languages of older civilizations’ were ‘well served by grammarians’, e.g., ‘Panini for Sanskrit’, ‘Dionysius for Greek’, ‘Priscian for Latin’, or ‘Al Khalil for Arabic’ (cf. Robins 1951) (P1 216; cf. 8.19, 58, 88). The trouble arose later when ‘Latin grammar was misapplied to an ever-increasing number of languages’, along with ‘Greek logic and metaphysics’ (P1 216; cf. 2.5; 3.50; 4.4; 5.24; 6.5; 9.25; 12.20f; 13.7, 16). Today, the ‘modern technician’ ‘finds traditional grammatical categories logically and philosophically pretentious, and a nuisance in practice’ -- a ‘medieval scholastic instrument’ ‘out of harmony with general scientific theory’ (TMS 87, 136). They ‘deal with’ ‘form and meaning’ ‘in the vaguest of logico-philosophical terms’, and even ‘some linguists follow’ this ‘method’ (TMS 136, P2 12; cf. 2.5; 3.4; 4.4f; 8.16; 12.7, 38, 64). Yet ‘traditional logic’ ‘shows no connection with or understanding of language, or ‘rational use of words and sentences in everyday life’ (TMS 104) (13.17).⁷

8.6 Firth also mistrusts ‘comparative, historical’ ‘philology’ with its affinities to ‘evolutionism’, ‘Darwinian’ ‘ideas’, ‘biological analogies’ and ‘the Romantic Reaction’ (TMS 147f, P1 16f; cf. P1 139; 2.5f, 13; 4.73f; 6.5; 8.15, 40; 12.20; 13.13). Like Sweet and Malinowski, he prefers the ‘analysis’ of ‘living languages’, at least until ‘descriptive linguistics’ and ‘functionalism’ enable a ‘reformulation of problems in comparative and historical work’ (P2 144f, P1 120, 218) (cf. 12.99). He also sees the ‘quest’ for the ‘origins’ of language as ‘largely futile’, citing the 1866 ‘Statutes’ of the ‘Societe de Linguistique’ against ‘accepting any communication concerning the origin of language or the creation of a universal language’ (TMS 144; cf. 8.3, 18; 2.47; 3.67; 12.17f). Instead of ‘going back’ and getting ‘further and further away from the habits we know and can observe’, we should ‘look for the origins of language in the way we learn and use it’ in ‘our everyday social life’ (TMS 189, 26) (cf. 9.12). Using quaint ‘reduplicative nicknames’ (‘bow-wow’, ‘pooh-pooh’, ‘ding-dong’, ‘yo-he-ho’, ‘ta-ta’), he glides over some ‘theories of the origin of speech’ from activities in the natural environment (TMS 25f, 141ff). Yet an account he favours is biological and Darwinian too: ‘the larynx’ first ‘“subserved the functions of locomotion, prehension, olifaction, and deglutition”‘ and the ‘voice’ was then developed for its ‘survival value’ in ‘leaving the hands and eyes free, travelling well’, and ‘conveying identity’ (TMS 145; cf. Negus 1949) (4.34).

8.7 In modern times, Firth sees another grave flaw in language study. Though ‘the distinction between “educated” and “uneducated” English dates’ from the ‘seventeenth century’ and its ‘grammarians’, ‘the Education Act of 1870’ was decisive, leading to ‘execrable’ ‘grammars for the young’ and to ‘much prejudice and difficulty of intercourse’, e.g., when the ‘schoolmaster's “educated” speech made children ashamed of the speech of their parents’ (TMS 195, P1 160) (cf. 4.40). In ‘traditional school grammars, the rules’ are ‘based on value judgments usually deprecatory’ and on ‘puritan’ ‘taboos’, e.g., against ‘using a preposition to end a sentence with’ (P2 120, 23; cf. 4.5).⁸ In America, the ‘Pure-English crusade’ greets ‘the living language’ with ‘“sneers and prohibitions”‘; ‘according to most writers on the subject, the speech habits of about 90% of the English-speaking world are bad’ (TMS 200). In return, the ‘artificiality’ of ‘“good English”‘ prevents ‘“schools and colleges from turning out pupils who can put their ideas into words with simplicity and intelligibility”‘ (TMS 201; cf. Mencken 1919; 4.85ff).

8.8 This grim situation provides the backdrop to Firth's ‘appeal for more disciplined modern linguistic studies’, including ‘grammar’, ‘in the English schools of the universities’ to accompany the ‘switch-over to science and technology just ahead of us’ (P2 117; cf. 4.6). ‘The problem of establishing a grammar of the main languages of life’ must ‘be dealt with at the level of science’, ‘by general or theoretical linguistics’ (P2 115). ‘In the journals of the future Institute of Language, competent technicians will give routine indications of any common or influential linguistic phenomena which are “definitely bogus”‘ (e.g., ‘the British transmogrification of the ablative absolute, and all the rest of the “bogus” grammar which teachers of classics impose’ on ‘vernacular languages’); ‘grammars’ and ‘language books will be “X-rayed” and the results stated with as little feeling as possible’ (TMS 105; cf. TMS 87).

8.9 ‘Clearing the litter’ of ‘generations of pedagogical mediocrities’ should provide ‘more wholesome surroundings’ for ‘children’ to do ‘linguistic exercises’ (TMS 105). ‘In schools’, ‘children should be shown how interesting it can be to talk in an orderly way’ ‘about their language as a vital part of their experience’ (P2 179, 120). Also, research can explore ‘the language of social control’ in ‘education’ and ‘apprenticeship’; ‘properly trained observers in nursery schools and all children's institutions’ can study ‘speech habits in formation’ (P2 179, TMS 151). And finally, we can ‘encourage open and natural use of local forms of speech’ and ‘literature’, not just the ‘Received English’ of ‘the elite of the public school class’ or a ‘purely negative’ English ‘free from unusual features’ -- a ‘strained form of speech’ that ‘masks social and local origin’ (TMS 200, 196f, 18, 200).

8.10 Meanwhile, Firth is ‘shocked to realize that we English, largely responsible for the future of the only real world language’ and ‘representatives of the civilization of all Europe in the four quarters of the globe, have up to the present made no adequate provision for the study of practical linguistic problems’ (TMS 211). ‘Some national provision should be made for more modern linguistic sciences on a scale commensurate with the wealth and position of Britain and America’ (TMS 138). By showing that ‘linguistics’ is a ‘more important social science’, we may ‘secure an endowment’ for an ‘Institute of Linguistic Research’ to address ‘educational, administrative, and social problems throughout the Empire’ (TMS 151, 211; cf. P1 172; 13.14).

8.11 Firth patriotically advocates language study because of ‘the vastness of our Empire’ with ‘so many religions, faiths and tongues’ (P2 144; TMS 138). ‘Great voyages of discovery’ led to ‘the widening of the linguistic horizon’, ‘the study of exotic alphabets’, and ‘World English’ (TMS 54, P1 103f). Even now, Firth sees ‘our young men and women’ ‘coming back from their voyages’, and predicts ‘another revival of learning’ (P1 103, 119, 141) -- a wistful hope for us in the 1990s, when British education has been brutally cut back by Conservative politicians. But we can still keenly appreciate the academic and political policy behind Firth's vision of ‘a new awareness everywhere of the powers and problems of speech and language’ (P1 141).

8.12 Firth hints that linguistics might reward investment by helping to unify the British Empire through language, and to ensure the worldwide preeminence of English. He proclaims that ‘the use of the English language today is the greatest social force in the whole world, and we in England should lead the way in training young people’, including ‘all foreigners who wish to join this active world fellowship’, ‘towards a critical understanding of language behaviour’ (TMS 137f). Otto Jespersen ‘found modern English the most advanced language in the history of mankind’; this ‘world language’ spoken by ‘the successful English and American peoples’ and vital to ‘the spread of European civilization and the culture of the white race’,⁹ should be ‘praised and used’ by ‘men of learning as well as men of affairs’ (TMS 209, 148). ‘For the sake of mankind it is to be hoped that English will drown the others; let all men of goodwill do their utmost to strengthen its service of mankind’ (TMS 54).

8.13 In contrast to the ‘older order of things’, when ‘knowledge and culture were the privilege of a small international elite using one international language of learning’ while ‘the vast masses were left in the dark’, Firth sees ‘the days of self-determination and popular culture’ dawning, when ‘every member of any considerable speech community should have the opportunity of cultivating’ the ‘mind’ ‘in his mother tongue’ (TMS 209f; cf. TMS 148f, P2 132; 8.4). But he finds the ‘social and cultural value’ of a language with fewer than 100,000 speakers ‘extremely doubtful’; ‘below 10,000 it almost ceases to be of any value outside the most primitive forms of group action’ (TMS 208). In contrast, English is said to be ‘spoken by 150 million people now [i.e. 1930]’ (swelling to ‘180 millions’ two pages later) and is ‘the official language of 540 millions’, a figure including all India so as to surpass Chinese with ‘430 millions’ (TMS 205, 207).¹⁰ These numbers fuel Firth's belief that ‘English is the only practicable world language’ and can be ‘taught in a normalized form’ (‘described by competent authorities’) ‘the world over as a second language’, being ‘easier to learn than French or German and much more useful’ (TMS 136, 200).¹¹

8.14 By ‘regarding language from a world point of view’, we can ‘carry out useful language work’ by ‘declassicalizing in both East and West’ and discarding the ‘distinction’ ‘between primitive and civilized’ ‘languages’ (P1 171, P2 135; cf. TMS 141). ‘To deal with the theory of language, the Western scholar must de-Europeanize himself’, and ‘the Englishman must de-Anglicize himself’ (P2 96) (cf. 2.32 3.5, 50; 13.42). Firth thus thinks it ‘all to the good’ to have the ‘chair of General Linguistics’ placed ‘at the School of Oriental and African Studies’, and hails the ‘enormous scope in the application of general linguistics’ ‘for the development of the free countries of Asia and Africa’ following ‘their rise’ (from colonialism!) (P1 171, P2 135). In ‘under-developed countries’, ‘nationalism leads to a longing for linguistic equality’, and ‘the leaders are quick to realize the value of linguistics’ ‘for their national aims’ (P2 131f, 135f).¹² Here, the ‘general linguist’ must ‘offer help and guidance’, though ‘the fields of research’ ‘cannot be fully exploited, even if all the linguists of the world were to unite’ (P2 135f).

8.15 Firth's national pride is conspicuous also in his attention to the work of past centuries. He lauds the ‘weighty contributions’ of ‘English linguists’, grammarians, rhetoricians, phoneticians, and orthographers (including shorthand inventors), and since the time of Elizabeth I or indeed since ‘Alfric's Latin Grammar’ ‘in English’ (P1 103, 100f). He cites such ‘pioneers’ as Thomas More, Thomas Wilson (1553), John Cheke, Thomas Smith (1568), John Hart (1569), Roger Ascham, William Bullokar (1580a, b), Timothe Bright (1588), Alexander Hume, Charles Butler (1634), Cave Beck (1657), John Wilkins (1668), William Holder (1669), John Wallis, George Dalgarno, Elisha Coles (1692), Thomas Gurney, John Byrom, William Blanchard, Isaac Pitman, William Jones, Walter Haddon, Richard Temple (1899a, b), Joseph Wright, the Bell family, and above all his idol Henry Sweet (our ‘pioneer leader’ and ‘greatest philologist’, ‘one of the cleverest thinkers on language’, etc.), with whom Firth ‘loved to be compared’ (Palmer) (P1 103, 100f, 110, P2 54, 137; cf. P1 168, 92-120, 166f; TMS 12, 63-66).¹³ Then too, Firth outdoes American theorists in citing early American linguists before Whitney: Samuel Haldeman, John Pickering, Peter Stephen DuPonceau, James Smithson, Alexander Bryan Johnson, and so on, plus grammarians and language planners like Benjamin Franklin, Noah Webster, and Lindley Murray; and he is quick to point out if they came from, or moved to, Great Britain (cf. P1 7, 157-166, 116-119).

8.16 Though his geographical and historical scope is expansive, Firth is uncertain about how broad his discipline should be. On the one hand, he feels ‘the linguistic sciences’ should seek ‘alliance’ with ‘the biological and social sciences’ and ‘develop proper semantic relationships’ with all ‘sciences of man’ (P1 143, 139) (cf. 9.7; 13.14). As ‘a social science’, ‘linguistics’ ‘is ahead of the others in theoretical formulation and technique of statement’; its ‘findings’ ‘are basic and must be carefully studied’ (P2 159, 189 cf. 8.33; 13.21). Remarking that ‘electronics has become a key subject’, and ‘mathematics and physics have always ruled us’,¹⁴ he sees ‘linguistics of the future becoming their opposite number’; ‘universities will encourage its study as one of the more austere disciplines fit to be ancillary’ to ‘sciences which promise us miraculous machines’ (P2 95). Besides, linguists ‘share’ an ‘interest in the meaning of meaning’ with ‘sociologists’, ‘biologically oriented psychologists’, and ‘philosophers’ in the ‘empiricist tradition’ of Locke, Hume, Moore, Russell, and Wittgenstein (P2 96) (12.97). Indeed, ‘during the next fifty years general linguistics may supplant a great deal of philosophy; the process has begun’, e.g. in Hjelmslev's Prolegomena (P1 168, P2 44) (13.16). Too, it should be ‘easier’ for ‘linguists’ to ‘acquire sufficient psychology and sociology’ than for ‘a psychologist or sociologist to acquire the necessary linguistic technique’ (P1 28) (cf. 12¹⁵).

8.17 On the other hand, Firth is gratified that ‘linguistics in Europe’ has ‘recently’ undergone ‘a revolution in status’ and ‘become an autonomous discipline’, not having its ‘point of departure in another science’ (P2 130, P1 177, 181, 190) (13.9-20).¹⁵ We incur a ‘great handicap’ by ‘depending’ on ‘prior disciplines’, such as ‘logic, rhetoric, philosophy’, ‘psychology, sociology’, ‘biology’, ‘pedagogy’, ‘metaphysics’, or ‘literary criticism’ (P2 130, P1 181, 191). Nor should ‘linguists’ ‘play second fiddle to cybernetics, communication theory, digital computers, speech machines, or telecommunications engineering’ (P2 130). And ‘our studies of speech and language’, as well as our ‘educational methodology, have been dominated far too much by psychology and logic’; ‘individual psychology’ ‘emphasizes’ ‘incommunicable’ ‘experience’, and ‘logic has given us bad grammar and taken the heart out of language’ (P1 186; cf. TMS 175; 4.5; 8.5, 7). So despite ‘some serious misapprehensions among linguists about modern logic’, it cannot ‘form an integral part of linguistics of any school’ (P1 217; cf. 2.5, 35, 84; 3.23; 5.7, 10, 41, 49, 54, 56; 8.36; 13.17). Besides, ‘philosophers are still unaware of the developments in linguistics during the last 30 years’, and their ‘analyses’ of ‘language’ and ‘meaning’ are ‘not linguistic’ (P2 70, 85; cf. 3.62, 4.4f, 51; 5.3, 5, 10, 13; 11.40; 13.16).

8.18 Such assertions cloud Firth's avowal that the ‘branches of linguistics cannot be seen in proper proportions and perspective’ without a ‘fundamental philosophy of language’ (TMS 3). Little can be made of playful, offhand remarks like ‘for his philosophy, the linguist need go no further than the second chapter of Genesis’, i.e., the ‘naming’ of ‘creation’ by the ‘magic power of the voice’, the more so as Firth elsewhere chides ‘the confusion of speech and life, race and language dating back to the book of Genesis’ (P1 35; TMS 76). Firth's allusions to the ‘magic’ of ‘language’ -- projecting from ‘linguistic’ ‘studies of magical word’ to the thesis that ‘language can be regarded as magic in the most general sense’, with ‘miraculous’ ‘creative functions’ (TMS 23, 32, 46, 113, 135; P1 185; P2 155) -- chiefly reflect his reaching for popular effect and his perplexity (sharpened by his anti-mentalism, 8.24, 41) about how the processing of language works.

8.19 A further check to philosophy arises if ‘it is not the task of linguistics to say what language is’ (P1 177) (cf. 9.1; 12.16, 39). ‘The techniques of linguistics have not been developed to deal with language in general human terms’ -- a limitation Firth relates to Saussure's ‘opinion that “le langage”‘ is ‘“inconnaissable”‘ (P1 190nf; cf. P2 110; 2.19). Firth can envision at most a ‘general physiology of utterance’, ‘perception’, ‘urges, and drives in our human nature’ (P1 191, 186), areas favoured for their generality and their amenability to behaviourism (cf. 8.22f, 25ff). We need ‘a general linguistic theory applicable to particular linguistic descriptions’ and ‘language problems’, ‘not a theory of universals for general linguistic description’ (P2 202, 190, 130f, P1 xii; cf. 7.19f; 8.60; 13.18).

8.20 Another peril Firth vehemently denounces is the ‘duality’ or ‘dualism’ between ‘mind and body’, ‘“signifiant et signifié”, expression and content’, ‘language and thought’, ‘thought’ and ‘expression’, ‘thought and word’, ‘idea and word’, etc., dating from Descartes and upheld by ‘Swiss, French, and Scandinavian linguists’ (TMS 20, 150, P1 19, 192, 227, P2 84, 90, 128, 203; cf. 2.25f; 13.10). Instead, Firth wants ‘general linguistics’ to adopt a ‘psychosomatic’ ‘approach’ to ‘mind and body taken together and acting in specific living conditions’ (P2 207). We must look to ‘the whole man thinking and acting as a whole, in association with his fellows’ (P1 19, 189, 225). Just as ‘the study of the whole man by biologists, anatomists, physiologists’, ‘neurologists, and pathologists is a commonplace of science’, ‘the linguist’ ‘must assume that normal linguistic behaviour as a whole is a meaningful effort’ for ‘maintaining appropriate patterns of life’ (P1 225; cf. 8.25ff, 47, 74; 13.13).

8.21 Firth thus insists that ‘the human body’ is ‘the primary field of human experience’ and ‘expression’, yet ‘continuous with the rest of the world’ (P2 199, 91) (cf. 4.8, 10, 13; 5.27, 42, 80; 11.12). Both ‘the body’ and ‘the world’ are ‘a set of structures and systems’ we can expect to discover in ‘the whole of our linguistic behaviour’ viewed as ‘a network of relations between people, things, and events’ (P2 90; cf. P1 143). Yet to say (with Whitehead) that ‘voiced-produced sound is organically rooted in living beings’ or (with Russell) that ‘meaning can only be understood if we treat language as a bodily habit’, raises problems if ‘most’ ‘organic processes’ are ‘intimate and secret’ (P2 206, 90, 199, TMS 150). ‘We need to know a good deal more of the action of the body from within, and especially of the nervous and endocrine systems’, if ‘human knowledge’ is ‘a function of that action’ (P1 143; cf. P1 142, 188; cf. 4.18f; 13.12).

8.22 All the same, Firth ‘views speech’ as ‘a bodily habit having a physical basis’ within ‘the central machinery for the control and coordination of behaviour’ (TMS 152f) (cf. 8.53, 69). As an ‘act’, ‘speaking’ points not only to ‘the functioning of the brain’, but to ‘localized speech centres with all their connected processes’, plus ‘movements of the face, head, arms, hands’, ‘legs’, ‘diaphragm’ and ‘abdomen’, and ‘general bodily gesture’ (TMS 152, 154; cf. 5.44).¹⁶ Some ‘influential schools of modern physiology and psychology’ hope for ‘a purely mechanical or materialistic explanation of all thought’ in terms of its ‘motor accompaniment’ (TMS 178). John B. Watson (1925) vowed ‘there is no such thing as thinking’, ‘only “inner speech” or the incipient activity of laryngeal and other speech processes’ (TMS 150, 179, P2 171; cf. 3.10; 4.9, 5.39; 13.12, 12²¹).¹⁷

8.23 However, Firth stops short of ‘Professor Pavlov of Leningrad’ by arguing that whereas ‘instinctive’ ‘habits’ ‘require no learning’, being ‘controlled by innate settings of the nervous system’, ‘all characteristically human habits’, including ‘speech’, ‘involve learning by experience’, ‘adjustment’, ‘experimental attunement, retention’ ‘recognition’, and ‘adaptability’ (TMS 180f) (cf. 7.30). Following the mood of the times, Firth may portray ‘words’ as ‘stimulus-response acts’, and ‘spoken sentences’ as ‘successions of directive stimuli’ to ‘evoke a suitable habitual response’ (TMS 175) (cf. 4.10-14; 5.67). But he stipulates that ‘habits’ ‘are based on’ ‘flexibility of response, substitutions, replacements’, and ‘variations’, whereby ‘intelligent behaviour’ can ‘adjust’ ‘to our environment’; ‘if intelligibility depended on a narrow reflex connection between speaking and hearing, we should all speak exactly alike and be no better than poultry’ (TMS 181, 23).¹⁸ Besides, ‘no two people pronounce exactly alike, and most’ ‘use more than one style’; ‘familiar sounds are constantly being made’ in ‘partly new contexts’ (TMS 181f; cf. 4.16; 5.38, 47; 8.77).

8.24 Firth's strong concern for ‘the bodily system, personality, and language through life’ (P1 188) leads him to capsule off the mental side, creating a fresh dualism. He applauds ‘Malinowski's warning’: ‘“all mental states” “postulated as occurrences within the private consciousness of man” are “outside the realm of science”‘; and ‘“there is nothing more dangerous than to imagine that language is a process running parallel and exactly corresponding to mental process” and “duplicates the mental reality”‘ (P2 158, 156) (cf. 5.10; 13.11). Hence, ‘general linguistics’ must not study ‘language as an instrument of thought’ or ‘an organ of the mind’ (P2 206; cf. P2 97ff) (but cf. 3.10ff; 5.69; 6.6; 7.10; 12.17ff, 22; 13.10, 14). Firth says ‘we do not deny the concept of mind, but we have no methodology or technique for studying it’ and ‘no technical language for mentalistic treatment’; nor should we ‘embrace materialism to avoid a foolish bogey of mentalism’, as Bloomfield did (P2 207; P1 192, 167 P2 175; cf. 4.8; 7.93). Nonetheless, Firth's horror of ‘taking refuge in mentalistic psychology’ (TMS 90) and his aversion to meanings apart from language -- he calls them ‘naked ideas’, as if their exposure were indecent (P2 75f, 78f, 81f, 85, 197) -- create important blind spots (cf. 8.50, 81ff).

8.25 He is even reluctant to ‘regard language as expressive or communicative’, lest he ‘imply it is an instrument of inner mental states’, ‘thoughts’, or ‘ideas’, which are ‘mysterious’ because ‘not observable’ (TMS 173, 135, P2 169f, 187). ‘Rather than as a countersign of thought’, Firth (with Malinowski and against Sweet, Whitney, and Hermann Paul) wants to ‘regard language as a mode of action’ (P2 148, TMS 150). ‘Language’ is thus ‘a way of doing things and getting things done’, of ‘behaving and making others behave’ ‘in relation to’ ‘surroundings and situations’ (P1 35, 31, 28f; cf. 4.88; 5.7ff, 12, 26f, 50; 8.20ff, 27, 46; 9.7ff; 11.5f, 11f). ‘By regarding words as acts, events, habits, we limit our inquiry to what is objective and observable in the group life’ (TMS 173) (13.12, 14, 45).

8.26 ‘A normal complete act of speech is a pattern of group behaviour’, ‘of common verbalizations’ of the ‘situational’ and ‘experiential contexts of the participants’ (TMS 173; cf. P1 35, TMS 135, 152). Admittedly, the ‘pattern’ is ‘without clearly defined boundaries’; ‘it is difficult to isolate and describe individual speech behaviour’. Yet there are ‘fine distinctions in speech behaviour determined by typical recurrent social situations’, wherein ‘locutions’ are ‘organs or functions’ (P1 75) (cf. 9.8, 40). Due to ‘contextual elimination’, ‘what you say raises the threshold against most of the language of your companion and leaves only a limited opening’ for a ‘likely range of responses’ (P1 32; cf. TMS 94). ‘Conversation in our everyday life’ is ‘narrowly conditioned’ by ‘culture’ and by ‘small speech groups, such as the family, caste, or class’; its ‘ritualistic give-and-take’ entails ‘grave social risks’ for ‘unexpected and highly individual’ ‘behaviour’ that seems ‘unusual’, ‘misdirected’, ‘tactless’, or ‘eccentric’ (P1 31, 75, TMS 181; cf. TMS 93). Moreover, ‘social requirements’ call for ‘clarity, decency, uniformity, and correctness of utterance’ (TMS 22). Indeed, ‘good manners require’ that ‘everyday speech’ be ‘full of banalities and cliches’; and for many ‘situations’ in ‘churches, law courts, or offices’, ‘conventionally fixed’ ‘words’ ‘bind people to a line of action’ (TMS 113, P1 30) (cf. 11.97).

8.27 Firth thus sees ‘the most universal forms of language behaviour’ in ‘routine service’ and ‘social ritual’, not in the ‘freedom’ of the ‘individual soul’; ‘even in literature’, ‘extreme’ ‘nonconformity is rare’ (TMS 94, 113, P1 28, 31) (cf. 3.38f, 64, 70). He proposes the term ‘tact’ for a ‘complex of manners which determines the use of fitting forms of language as functional elements of a social situation’; and the term ‘set’ for a ‘general pattern of behaviour’ ‘belonging to a social group’ or ‘type’, including ‘instincts’ (‘sex, hunger, fear, anger’), urges, sentiments, interests, abilities’, ‘feelings’, and ‘curiosity’, plus a ‘sense of order and system’, ‘fellowship’, ‘superiority, inferiority, snobbery’, ‘obligation’, ‘licence’, ‘submission’, and ‘domination’ (TMS 17, 89f, 95, 100; cf. 8.19). The ‘language behaviour’ ‘observed in the actual context of situation’ ‘may be regarded’ as a ‘manifestation’ of the ‘sets’, which ‘tune themselves automatically to ‘link selected input with appropriate output’ (TMS 93). Still, the ‘sense of order’ -- you may ‘call it’ ‘reason, thought, intelligence, logic, genius, insight, inspiration, or just craftsmanship’ -- ‘is always in peril of being overruled’ by ‘primitive feelings’ ‘we share with the animals’ (TMS 100, 95).

8.28 If both ‘language and personality are built into the body’, and ‘the organization of personality’ ‘depends on the built-in potentialities of language’, then ‘linguistics’ must address ‘the key notion’ of ‘personality’, whose ‘basic principles’ are its ‘unity, identity, and continuity’ within ‘the social process’ and ‘the creative effort and effect of speech’ (P1 143 184, 141, P2 13f) (cf. 3.70ff; 5.7, 26; 6.2, 54; 8.40, 43, 81; 9.14, 71; 11.18, 66; 12.58). Moving ‘a long way from Saussure's mechanical structuralism’, in which ‘the speech’ of ‘the underdog speaking subject’ ‘was not the “integral and concrete object of linguistics”‘, Firth stresses ‘the study of persons, personality and language’ as ‘vectors of the continuity of repetitions in the social process and the persistence of personal forces’ (P1 183; cf. 2.20; 3.1; 4.10; 9.7). Due to the ‘close association between personality and social structure’, he favours ‘sociology’ over ‘individual psychology’ (P1 185f; cf. 9.7; 13.14). ‘Linguistics’ ‘is mainly interested in persons and personalities as active participators in the creation and maintenance of cultural values’, rather than as ‘separate natural entities in their psycho-biological characters’. Hence, ‘the study of spoken language’ should ‘stress the study of specific persons’ over ‘the collection of haphazard and colloquial oddments’ from ‘speakers at random’ (P2 32).¹⁹

8.29 ‘The human being’ is therefore to be ‘regarded’ not ‘as an individual’, but as a ‘person’ ‘acting in his many social roles’, whose ‘interaction’ is ‘a conservative force in personality’ and ‘society’ (P2 207, P1 28f). ‘The relevant forms of language’ are ‘the lines of the leading roles’, which ‘interlock’ without ‘conflict or serious disharmony’ in ‘an integrated personality’, favouring ‘social responsibility and stability’ (P2 207; cf. P1 186). Firth chooses ‘the terms idiolect for ‘a form of language used between two personalities chiefly in one of their personal roles’ versus ‘monolect’ for a form limited to one person, e.g. by ‘language disorder’ (P2 209). Hence, he ‘recommends that more attention be given to linguistics by psychoanalysts and psychiatrists’, just as ‘general linguistics’ should ‘place more emphasis on our activities, drives, needs, desires, and tendencies of the body than on mechanisms and reflexes’, although it ‘recognizes’ them ‘indirectly’ (P2 209, P1 143, 225; cf. 8.19).²⁰

8.30 The breadth of concerns outlined so far should indicate why ‘for any given language’ ‘no coherent system’ ‘can handle and state all the facts’, and ‘linguistics must be polysystemic’ (P2 24, P1 121, P2 43, 200; 9.19). Firth therefore repudiates ‘the monosystemic principle’ ‘stated by Meillet (“Chaque langue forme une systeme ou tout se tient”’) as a pretext for ‘static structural formalism’ and ‘mechanical materialism in linguistics’; and rejects Saussure's division between ‘“langue” and “parole”‘ (P1 180f, 121, 144, P2 28, 41, 127, 139f; cf.2.20; 5.7; 6.33, 46; 7.12; 8.30; 9.5; 12.12, 26, 47, 55, 67; 13.36).²¹ Firth hopes for a ‘synthesis of contemporary theories’: ‘we are all right’, not merely ‘dogmatic interpretations of Saussure’ (P2 24, i.a.). ‘Polysystemic’ ‘hypotheses’ may not render ‘problems easier’ than ‘the monosystemic analysis based on a paradigmatic technique of opposition’, but may render ‘the highly complex patterns of language clearer’ within ‘the plurality of systems’ that are ‘not necessarily linear’ nor related to ‘successive fractions or segments of the time-track of instances of speech’ (P1 137, cf. P1 147) (cf. 13.35).

8.31 In this connection, Firth takes pains to deny being ‘a “structuralist”‘, a school he traces back to Baudouin de Courtenay, Saussure, and Meillet (P2 145, 41). ‘Structuralist’ work ‘forms only one part’ of ‘structural linguistics’ that from his ‘point of view’ ‘aims at employing all technical resources systematically for multiple statements of meaning in the appropriate linguistic terms’ (P2 145, 44, 50, 3; cf. 8.45, 48f). In contrast, ‘structuralism’ ‘emphasizes segmentation and phonemics’ and ‘excludes meaning’ (P2 47; cf. P2 44, 48, 129; cf. 4.14ff; 5.61; 7.56ff; 13.27). Originally ‘an established technique for reducing languages to writing’ ‘in ethnographic studies of the American Indian’, ‘phonemics for its own sake’ became ‘a theoretical discipline’ -- a trend ‘like generalizing pure mathematics from practical arithmetic’ (P2 129). Moreover, structuralism seeks ‘a linguistic mathematics or a completely axiomatized science’, which in Firth's estimate ‘will not be found workable as a truly empirical science’ but will ‘become a dead technical language’ (P2 47; cf. P2 43) (cf. 13.15).²² Nor is he deeply impressed by the ‘use of logical and statistical theories’, or with displays of ‘postulates and axioms’ to announce ‘a more scientific methodology akin to mathematics’ (P2 129, 43; cf. 8.17, 55; 13.17).

8.32 As a ‘first principle’, Firth recommends ‘distinguishing between structure and system’ (P2 186). ‘Structure’ is ‘syntagmatic’ and ‘horizontal’, whereas ‘system’ is ‘paradigmatic’ and ‘vertical’ (P2 186, 200) (cf. 2.30, 56, 65; 6.34; 9.3; 12.71). ‘Since systems furnish values for elements of structure and since the ordering of systems depends on structure’, ‘the exponents of elements of structure and of terms in systems are always consistent’, though not of the same ‘order’ -- ‘the term exponent’ referring here to the actual ‘shape of words or parts of words’ (P2 183f; cf. 8.53, 59, 68f; 12.86).²³ Against the Americans, Firth rejects any account in terms of ‘segments in any sense’; ‘elements of structure’ ‘share a mutual expectancy in an order which is not merely a sequence’ (P2 186, 200f; cf. 8.52, 60, 64, 69). He further postulates such an ‘expectancy not only between’ ‘elements of discourse’, but also between ‘words and the surrounding living space’ (P2 206). The fact that expectancy is a mentalistic notion does not disturb him, since he usually assigns it not to people but to language ‘elements’ like ‘words and sentences’ (P1 195, P2 69, 181, 186, 200, 206) (cf. 8.64).

8.33 As we might expect, Firth ‘ventures to think linguistics is a group of related techniques for handling language events’ (P1 181, 190). For ‘systematic empirical analysis’, ‘descriptive linguistics must be practical’; ‘its abstractions, fictions, inventions’ must be ‘designed to handle instances of speech, spoken or written’ (P1 173) (cf. 9.4; 13.36). ‘A theory derives its usefulness and validity from the aggregate of experience to which it must continually refer’ (P2 168). Yet just as ‘theories are inventions’ or ‘constructions’, ‘four-fifths of linguistics, including even experimental phonetics, is invention rather than discovery’ (P2 124, P1 173) (cf. Panconcelli-Calzia 1947). ‘Systematics’ are ‘ordered schematic constructs, frames of reference’ or ‘scaffolding’ intended to ‘cover’ a ‘field of phenomena’; they ‘have no ontological status’, ‘being, or existence’; ‘they are neither immanent nor transcendent, but just language turned back on itself’ (P1 181, 190, 121, 147, P2 124; cf. 8.23; 9.27; 13.48). ‘The descriptive linguist does not work in the universe of discourse concerned with “reality”‘, nor ask ‘whether his isolates can be said to “have an existence”‘ (P2 155f; cf. P2 154; 8.44; 5.12f, 28f; 6.12; 13.57). At best, such ‘fictions’ as ‘speech, language’, ‘tact’, ‘set, and context of situation’ ‘have a certain practical value’ (TMS 135).

8.34 Since ‘we are all participants in those activities which linguistics sets out to study’, and ‘all linguists rely on common human experience’ to ‘make abstractions’, ‘linguistics is reflexive and introvert’ (P2 169, 8, P1 121, 147) (cf. 1.8f; 12.12; 13.48). ‘Each scholar makes his own selection and grouping of facts determined by his attitudes’, ‘theories’, and ‘experience of reality’, and his ‘statements’ must ‘be referred to personal and social conditions’ (P2 29) (cf. 13.1, 36). So we should be ‘constantly mindful of the different levels of abstraction’ (P1 173). Reciprocally, Firth insists that ‘there are no scientific facts until they are stated’ ‘in technical language’ within ‘a system of related statements all arising from a theory and application’ (P2 30, 43, 154, 199) (cf. 13.48). He suggests we ‘regard’ ‘facts’ as ‘myths’ ‘we believe’ and ‘live with’, and quotes Goethe's epigram: ‘“the highest state would be to grasp that all facticity is already theory”‘ (P2 156, 146, m.t.).

8.35 For clarification, Firth distinguishes three ‘methods of stating linguistic facts’: (1) ‘language under description’ (‘exemplified by texts’), (2) ‘language of description’ (‘technical terms’, ‘notation’, etc.), and (3) ‘language of translation’ (P2 49, 87, 98, 112, 149, 158, 202). We shall survey them in turn. The language of description is essential because linguistics can contribute to ‘the progressive standardization of the universal languages of science’ and the ‘internationalizing of terms’ (TMS 67, 121, 210). ‘Sciences’ try to ‘frame international languages to serve their special needs’, and Firth thinks it ‘desirable’ ‘that an international technical language be developed in English for the use of linguists all over the world’, even if he does ‘not believe very much in conferences on standard terminology’, nor in ‘the internationality and universality’ of ‘linguistics’ as ‘common united knowledge in a variety of languages’ (TMS 71, P1 140, P2 28ff; cf. TMS 106). ‘Scholars’ in some fields may ‘agree that a number of alternative theories can be regarded as special cases’ of one ‘comprehensive theory’, but ‘linguistics’ is ‘far from’ ‘this happy state’ (P2 29) (cf. 5.9; 13.3). Instead, ‘linguists’ ‘have always disagreed most about terminology and nomenclature’ in their own ‘technical’ language (TMS 71, P2 83).

8.36 This state of affairs is aggravating, because ‘linguists should be the first to control, direct, and specialize almost every word they write in linguistic analysis’ and should remain ‘language-conscious at all levels’ (P2 34). ‘Scientific terminology is in no sense self-explanatory’, but ‘relative and functional’ (P2 120) (cf. 13.48). ‘We must have orderly language to discuss language, which is obviously based on ordered relations’ (P2 120). ‘The accurate formal description’ of ‘the constituents of a language’ ‘demands a highly specialized technique’ (TMS 87). ‘The more we take humanism out of linguistics the more we must’ ‘examine’ ‘our languages and techniques of statement’ to see if they are ‘proper’ ‘for a science’ (P2 34). In ‘such a science’, we should ‘state our results not only for one another but for all who may need them’ (cf. 2.88; 3.2; 13.62).

8.37 Firth himself can't decide between new terms and old ones.²⁴ Hjelmslev is lauded for his (probably ‘unsuccessful’) ‘attempt’ to create a ‘terminology’ within ‘a system of thought’ and ‘a rigid framework in which to test a special language for linguistic analysis’, so as to ‘emancipate’ us ‘from the handicap of the common-sense idiom and “self-explanatory” nomenclature’ (P2 46) (cf. 6.3, 55, 59). But work ‘bristling with neologisms’ brings ‘discredit’ to ‘modern linguistics’ (P1 140). The Americans are scolded both for creating ‘a set of unattractive new terms’ and for using ‘a large number of traditional terms’ like ‘“possessive case”‘ ‘without definition’ (P2 119). ‘The further we go with modern studies’, ‘the more ridiculous our traditional grammatical apparatus will become’, ‘due to its naivete and obvious incompleteness and inadequacy both in formal description and in dealing with meaning’ (TMS 87, P2 188; cf. 7.4; 8.5, 7, 58; 13.5, 7).

8.38 Although it would be ‘foolish to abandon’ it, the ‘nomenclature’ of the last ‘two thousand years’ should be ‘checked and sorted out’, and ‘fitted into an entirely serviceable technical apparatus for linguistic analysis and statement in keeping with the advances of linguistic theory’ (P2 189) (cf. 2.6; 6.49; 7.4, 75; 12.41, 88; 13.7). ‘Most of the older definitions’ of ‘“system”‘, for instance, including Saussure's, ‘need overhauling in the light of contemporary science’ (P1 143; cf. 2.26f, 36; 6.34ff; 8.30, 32, 53). For the future, Firth calls for ‘a systematic study of the languages of linguistics’ by ‘applying the techniques of semantics, both historical and descriptive, to the language used about language’ and to ‘the conceptual framework and systems of ideas’ wherein ‘our own technical terms’ ‘function’ as ‘focal or pivotal’ ‘key words’ (P1 139, 141, P2 98) (cf. 13.36). Also, we should ‘welcome new systems of linguistic thought with their terminology’ as a ‘radical criticism’ of ours (P1 141).

8.39 We should bear in mind too that the ‘technical restricted languages’ in which ‘the empirical data of such sciences as linguistics are usually stated’ ‘involve indeterminacy’, since the ‘terms’ can also appear in ‘common usage in general language’ (P2 46) (13.69). ‘Linguistics which does not fully recognize this’ cannot be ‘applied to the study of language in society’. Although ‘the general national language’ imposes some ‘epistemological conditions of scholarship’ ‘in ‘modern linguistics’, ‘the meaning of a technical term’ ‘cannot be derived or guessed at from the meaning of the word in ordinary language’ (P2 27, 169).

8.40 For ‘the new approach’ Firth envisions, a compromise is proffered: ‘many of the traditional terms survive, but their meanings are determined by the new contexts in which they are used’ (P2 119) (cf. 9.29; 13.48).The ‘pivotal terms’ ‘are given their “meaning” by the restricted language of the theory’ and its ‘applications’ (P2 169). In a chapter whose abstract announces ‘a rectification of terms’, terse explications are offered for ‘language’, ‘speech event’, ‘nature’, ‘nurture’, ‘system’, ‘phoneme’, and ‘feature’ (P1 139, 142-46). But other terms like ‘philology’ and ‘personality’ are critiqued by consulting Dr. Johnson and the Oxford English Dictionary (P1 141f, 183f) -- hardly ‘new’ or ‘radical’ sources.

8.41 Ultimately, ‘abstract linguistics’ gets its ‘justification’ when ‘the linguist’ ‘finally proves his theory by a renewal of connection with the processes and patterns of life’ and ‘experience’ (P2 19; cf. P1 xii, P2 17, 19, 24, 129, 154, 168, 175, 184f, 190ff) (13.57). Firth presents his own ‘monistic approach’ as the means whereby those ‘processes and patterns’ ‘can be generalized in contexts of situation’ (P1 226, P2 24, 90, 169).²⁵ Because ‘a speech event’ is an ‘expression of the language system from which it arises and to which it is referred’, ‘we can only arrive at some understanding of how’ ‘language’ ‘works’ if we ‘take our facts from speech sequences’ ‘operating in contexts of situation which are typical, recurrent, and repeatedly observable’ (P1 144, 35) (5.10). To avoid invoking ‘mental processes’, Firth suggests that any ‘memory context or causal context’ must be ‘linked up with the observable situation’ (P1 19; cf. 8.24). ‘A stated series of contexts of situation’ may thus contribute to ‘a theory of reciprocal comprehension, level by level, stage by stage’ (P2 200; cf. 5.65; 8.48).

8.42 To ‘make sure of the sociological component’, each ‘context’ should in turn ‘be placed in categories’ ‘within the wider context of culture’ (P1 182, 35; cf. 8.48; 13.62). Though ‘situations are infinitely various’ and cannot be ‘strictly classed’ within ‘hard and fast lines’ (4.13, 31, 61; 5.25, 28; 13.26, 40), Firth sees ‘great possibilities for research and experiment’: ‘contexts can be grouped into types of usage’ and ‘social categories’ like ‘common, colloquial, slang, literary, technical, scientific’ and so on, applying ‘the principle of relative frequency’ (P1 28, P2 177). Or, we can ‘refer contexts to a variety of known frameworks of a more general character’: ‘economic, religious and other social structures’; ‘types of linguistic discourse such as monologue, choric’, ‘narrative, recitation, explanation, exposition’; ‘personal interchanges’, varying with ‘the number, age, and sex of the participants’; ‘speech functions’ such as ‘address’, ‘greetings’, ‘direction’, ‘control’, ‘drills’, ‘orders’, ‘flattery, blessing, cursing, praise, blame, concealment’, ‘deception, social pressure and constraint’, and ‘verbal contracts of all kinds’ (P2 178f, TMS 111, P1 30f) (cf. 9.71).

8.43 ‘A situation is a patterned process conceived as a’ ‘dynamic and creative’ ‘complex activity with internal relations between its various factors (TMS 110f) (cf. 4.16; 9.11). ‘The ‘relations’ in a ‘situation’ may be among ‘the participants’ (as ‘persons’ and ‘personalities’), their ‘verbal’ and nonverbal actions’, and the ‘effects’ of these, plus ‘relevant objects’ and ‘events’ (P2 108, 148, 155, 173, 177, P1 182). ‘The text’ ‘is seen in relation to the nonverbal constituents and the total effective or creative result’ (P1 18n). Firth thus shares Pike's concern for ‘relations’ ‘between elements of linguistic structure and nonverbal constituents of the situation’ (P2 203, 148, 173, 177) (cf. 5.8; 11.7, 9, 86). Indeed, ‘references to the nonverbal constituents’ may be ‘essential’ for describing ‘formal linguistic characteristics stated as criteria for setting up parts of speech or word classes’, e.g., ‘nominal and verbal categories’ (P2 187) (cf. 9.2; 13.40).

8.44 All the same, Firth leaves no doubt that ‘in contexts of situation’, ‘the text is the main concern of the linguist’ (P2 24, 90, 173) (cf. 13.39). ‘All texts’ ‘in modern spoken languages’ are considered to carry the implication of utterance’ and are ‘referred to typical participants in some generalized context of situation’ (P2 201 13, 98, 123, P1 220, 226; cf. P2 30f, 175; 8.72). We should thus give ‘due attention to the form of discourse’ and ‘the style and tempo of utterance’ (P2 32; cf. 8.65, 69). The ‘attested language text’ should be ‘duly recorded’ and ‘abstracted from the matrix of experience’, which Firth likes to call, after Whitehead, ‘the mush of general goings on’ (P2 199f, 99, P1 187) (hardly suggesting order or structure). ‘The linguist’ then retains the ‘selected features or elements of the cultural matrix of the texts’ needed to ‘set up’ ‘formal contexts of situation’ (P2 200). Although ‘it may not be ‘possible or desirable to present the whole of the materials collected during the observation period’, some ‘“corpus”‘ is ‘essential’ (P2 32). This emphasis on ‘texts’ (see also P1 xi, 75, 145, 192, P2 13, 18f, 69, 85, 97, 108, 121, 140, 145, 169, 177, 181, 187, 190, 196) is shared by Pike, Hjelmslev, Halliday, van Dijk and Kintsch, and Hartmann (cf. 5.5, 15; 6.37; 9.1, 3, 8, 16, 41f, 107, 9¹⁹; 11.1f; 13.39).

8.45 Less widely shared is Firth's demand for a ‘situational approach’ derived by ‘general theoretical abstraction with no trace of “realism”‘ (P2 154; cf. 6.12; 8.33; 13.57).²⁶ ‘The context of situation’ would be a ‘schematic construct to be applied to language events’ as ‘technical abstractions from utterances and occurrences’ (P2 154, 175f, P1 144, 182; cf. P2 200). ‘Since science deals with large average effects’ via ‘observation’, we should ‘generalize typical texts or pieces of speech in generalized contexts of situation’ (P2 13). ‘We study the flux of experience and suppress most of the environmental coordination’, looking for ‘instances of the general categories of schematic constructs’ (P2 16). The ‘elements of the situation, including the text, are abstractions from experience and are not’ ‘embedded in it, except perhaps in an applied scientific sense’ (P2 154). Again repudiating mentalism, Firth excludes from ‘the concern of linguistic science’ ‘the intention of a particular person in a particular instance of speech’ (P2 16) (as Saussure excluded ‘the will of speakers’, 2.20, 44).

8.46 Despite his behavioural outlook, however, Firth is adamant that ‘descriptive’ or ‘structural linguistics’ ‘deals with meaning throughout the whole range of the discipline’ and ‘at all levels of analysis’; ‘meaning must be included as a fundamental assumption’ and ‘main concern’ (P2 50, 160, P1 190; cf. P1 xi, P2 145, TMS 86). ‘Linguistics’ can attain ‘no unity’ or ‘synthesis’ ‘unless we all turn’ to ‘the “second front”‘ of ‘linguistic meaning’ (Jakobson's term) (P2 48, 50, 74, 159f). The warning that ‘linguistics without meaning is sterile’ is addressed to American linguists like Bloch, Trager, and Harris, who claimed to ‘exclude meaning’ (P2 160, 47f, 85, 117, P1 227) (5.61; 7.56; 13.27). Yet Daniel Jones, whom Firth esteems, also avowed that ‘meanings’ don't ‘enter into the definition of a phoneme’ (P2 48; cf. 2.70; 4.26, 28; 6.43; 7²⁷). And, Firth himself is ‘convinced of the desirability of separating semantics’ ‘from grammar of the technical and formal kind’ (P1 6, 16; cf. 8.62; 13.54). But without more illustration, I can't tell what he expects from ‘a strictly formal study of meaning’ ‘in strictly linguistic terms’ (P2 160, 169, 81, 97, 176, P1 x).

8.47 Predictably, Firth's idea of a proper ‘“semantics”‘ is a ‘situational and experiential study’ of ‘contexts of situation’ ‘along sociological lines’, in which ‘linguistics accepts’ ‘language texts as related’ to ‘the “meaning” of life’ (P1 27, 16, P2 169, TMS 113, 184; cf. P2 82, 206; 8.41; 13.14).²⁷ ‘Meaning is best regarded’ ‘as a complex of relations between component terms of a context of situation’ (TMS 110f, 174; cf. P1 183). Only a ‘contextual theory of meaning employs abstractions which enable us to handle language in the interrelated processes of personal and social life’ (P2 14). Against the ‘logician’ who ‘treats words and sentences as if they could somehow have meanings in and by themselves’, Firth agrees with Wittgenstein that ‘words’ ‘mean what they do’ and ‘the meaning’ ‘lies in their use’ (TMS 110, P2 138, 162; cf. 4.78; 12.42, 66; 13.36). Seeing ‘meaning’ ‘deeply embedded in the living processes of persons maintaining themselves in society’ defies ‘intellectuals’ who emphasize ‘only the symbolic use of words’ over ‘“evocative” or “affective” language’ (P2 13; TMS 176) (cf. 9.15). For Firth, most ‘abstract words are based on other words integrated in ordinary social behaviour’; the truly ‘primary words’ are those ‘standardized’ by ‘convergence of action in some social group’ (TMS 176; cf. 3.45; 4.24; 5.66). He wonders if ‘the promotion’ and ‘maintenance of communion of feeling is perhaps four-fifths of all talk’ -- even at ‘conferences and congresses’ in ‘science’ (TMS 112; cf. TMS 110, 175).

8.48 Because ‘the statement of meaning’ for ‘whole texts’ is a ‘vast subject’, we must ‘split up the problem’ (P2 108; cf. P1 10,18f). One recourse Firth airs but rejects is to distinguish ‘denotation’ as ‘primary meaning (except perhaps’ ‘as highest-frequency meaning’) from ‘connotation’ or ‘secondary meaning’ or ‘connotation’ (P1 10f) (cf. 4.21f, 25; 6.52ff; 12.19). He especially rejects the idea, put forth by Trench (1832) and Skeat (1887), that the ‘primary’ or ‘true meaning’ is the ‘original’ or ‘etymological’ meaning (P1 8, 10f, P2 85, 121, 149). Such an idea emerges only when ‘scholars’ ‘study change’ of meaning and look for a ‘seminal meaning’ as the ‘ultimate origin’ (P1 19, 9). Firth places his own hopes on ‘generalization’ as a means to ‘avoid the appalling consequences of the continuous change of content in all expressions of a living language and of the belief that meaning can only be real in individual instances of human invention’ (P2 118; cf. 4.16; 8.19; 12.67f). With Stocklein (1898) and Sperber (1923), Firth demurs also that ‘a change of meaning is not’ ‘in the single word’, but in a ‘functioning context’ (P1 13).

8.49 Instead of these recourses, Firth's ‘central proposal’ ‘is to split up meaning’ ‘into a series of component functions’ (P2 173, P1 19). ‘The phonetic’ is ‘a minor function’, whereas ‘lexical, morphological, syntactic’ and ‘situational’ are ‘major functions’ (P1 20, 24, 33, 35, 37, 40f, 48f, P2 174).²⁸ To ensure an ‘analysis in terms of linguistics, we first accept language events as integral in experience’, ‘whole’, ‘repetitive and interconnected’, and then ‘apply theoretical schemata’ and make ‘statements’ in terms of ‘structures and systems at a number of levels of analysis’ (P2 176, 97). For this purpose, he envisions a ‘spectrum of linguistic analysis whereby’ ‘the total meaning of a text in situation’ is ‘dispersed at a series of levels’ (P2 92, 33, 81f, 97, 108, 110, 112, 118, 124, 174, 200f, P1 xi, 19, 24, 183, 192, 195, 220) (cf. 8.52, 61f, 84). This ‘“spectrum” analysis makes sure of the social reality of the data’ ‘before breaking down the total meaningful intention’ (P1 170f).

8.50 ‘Meaning’ thus becomes a term ‘for the whole complex of functions which a linguistic form may have’ (P1 33, P2 174; cf. P1 7; 8.64). ‘Semantic study’ becomes the place where ‘the work’ of ‘the phonetician, grammarian, and lexicographer’ is ‘integrated’ (P1 27). Hence, ‘semantics’ can proceed only if ‘phonetics’, ‘morphology’ and ‘syntax’ ‘are sound’ (P2 197, 33; P1 18f, 23, 28, 75) (cf. 7.57). This rationale allows his discussions of meaning to keep turning to those other levels. One paper title ‘Further Studies in Semantics’ got changed into ‘Sounds and Prosodies’ (P1 123). In another paper called ‘The Techniques of Semantics’ (P1 7-33), the only treatment of semantics in the usual sense is an attack on approaches Firth dislikes (e.g. Ogden and Richards, who appeal to ‘relations in the mind’); otherwise, the paper deals with past mentions of the term ‘semantics’ (e.g. in ‘the Society's Dictionary’), historical principles, morphology, and phonetics. Finally, the ‘context of situation’ gets referred to ‘sociological’ issues like ‘social roles’ and ‘customs’, or ‘cultural heritage’. Such papers are emblematic: Firth keeps heralding a new ‘semantics’ for making ‘statements of meaning’, yet instead of making such statements, he ‘disperses meaning’ to other levels. He never really proves that ‘the accumulation of results at various levels adds up to a considerable sum of partial meanings in terms of linguistics without recourse to any underlying ideas’; or that if ‘linguistic analysis’ ‘states the structures it finds both in the text and the context’, these ‘statements’ ‘then contribute to the statements of meaning’ (P2 197, 17).

8.51 Nor is Firth's overall scheme of levels terribly clear. Although ‘level’ is not a term in the early books, the two later ones mention some twenty-four ‘levels’ in various listings.²⁹ Most frequently named are: ‘phonetics’, ‘phonology’, ‘grammar’, ‘morphology’ (or ‘morphematics’), ‘syntax’ (or ‘syntagmatics’), ‘stylistics’, ‘situation’ (or ‘context of situation’), and ‘collocation’; only occasional reference is made to ‘pronunciation’, ‘phonaesthetic’, ‘semantic’, ‘lexical’, ‘vocabulary’, ‘word formation’, and ‘colligation’; and ‘prosodic’, ‘graphematic’, ‘spelling’, ‘sociological’, ‘phrase formation’, ‘word description’, ‘word isolate’, ‘etymology’, and ‘glossaries’ are termed ‘levels’ only once each (P1 xi, 24f, 171, 192, 197, 206, P2 16, 18, 30, 40, 91f, 97, 99, 106, 110f, 113, 118, 124, 127, 147, 149, 154, 175f, 181ff, 188, 195, 200ff, 208). Sometimes too, ‘modes’ is used in a very similar sense (P1 192, 198f, P2 33, 82, 110). If ‘the levels of analysis’ are ‘constantly increasing in number and specialization’ (P2 82), Firth's work is a good foretaste.

8.52 Comments on the ‘series of levels’ indicate that the ‘higher levels’ are closer to ‘culture’, ‘context’, and ‘situation’, and the ‘lower’ to ‘phonology’ and ‘phonetics’ (P1 201f, P2 33, 175f). Firth may say that analysis can go either way; or that ‘the descending procedure is the right one’, ‘the total complex’ of ‘the higher levels’ being ‘a first postulate’; or he may ‘adopt’ the ascendant one to avoid or postpone dealing with ‘ideas’, ‘concepts’, or ‘thoughts, and may not get to the higher ones at all, as in his treatment of Swinburne (P1 192, P2 175, P1 171, 198ff, 202f; cf. 8.83). In favour of his ‘prosodic approach’ (8.64ff), he argues that ‘a theory of analysis dispersed at a series of levels must require synthesis’ and ‘congruence of levels’, and (against Pike) that ‘all levels are mutually requisite’; yet elsewhere that the ‘analytic dispersion does not imply that any level constitutes a formal prerequisite of any other’: ‘the levels’ ‘are only connected in that the resulting statements relate to the same language text’; and ‘the levels’ must be ‘congruent and complementary in synthesis on renewal of connection with experience’ (P2 202, 111, 30, 192, 176f; cf. 8.40, 54).

8.53 Some of this perplexity reflects the difficulties of ‘determining what are the units of speech’ (TMS 182). ‘General opinion’ points to ‘words’, which, particularly in ‘literate societies’, are often ‘institutionalized’ (TMS 182f, P2 155, P1 122, xi). Too, ‘word analysis is as ancient as writing’ and has produced ‘the alphabet’ (P1 122, TMS 33; cf. 4.42, 63; 6.50; 8.73). Thus, for certain ‘purposes’, ‘words’ can be viewed as the ‘principal isolates’ within ‘texts’ (P1 122, xi). An ‘emphasis on the word as a unit is a useful corrective’ for ‘over-segmentation and fragmentation’ (P2 40, i.r.; cf. 8.31f). Sometimes, Firth imagines ‘words staring you in the face from the text’ (inspired maybe by Wittgenstein's idea that ‘a word in company’ has a ‘physiognomy’) (P1 xii, P2 179, 186). Other times, he ‘regards’ ‘words’ as ‘bodily’ or ‘habitual acts’, ‘events, habits’ (TMS 173, 149, 181; cf. 8.6, 21ff, 25). In that mood, he pictures ‘families of linked words’ arising from ‘related habits’; a word with ‘fewer derivatives and analogues’ has a ‘weaker background of bodily habit’ (TMS 183).

8.54 Yet Firth sees little prospect of an ‘acceptable definition of the word’ ‘in general human terms’ (P1 191n) (cf. 2.18; 3.31; 4.54, 60; 5.53; 6.23; 8.54; 13.29). He appears to have three motives for this reservation. First, words are not sufficiently theoretical entities. In grammar and syntax, units and structures involve ‘categories’, ‘relations’, or ‘abstractions’ rather than ‘words’ (P1 180, P2 112f, 121f, 152, 181, 186, 203; cf. 8.79). Here at least Firth agrees with the ‘strict Saussurian doctrine’ of ‘structural formalism’: ‘language’ is ‘a system of signs placed in categories’, i.e., of ‘differential values, not of concrete and positive terms’ (P1 180; 2.26). Second, words usually appear within larger units. Already for ‘the Ancient Hindus’, ‘“a word had no existence detached from a sentence”‘, and ‘“resolving a sentence into parts”‘ was ‘“a fanciful procedure”‘ (TMS 15) (cf. 9¹³). Moreover, ‘since in many languages the exponents of the grammatical categories may not be words’, ‘syntactical analysis must generalize beyond the level of the word isolate’; in fact, ‘some scholars’ believe there are ‘wordless languages’ (P2 182, 122).³⁰ Third, words are not satisfactory for dealing with meaning. Firth deplores ‘the naive approach’ which ‘regards’ ‘the meanings of words’ as ‘immanent essences or detachable ideas’ (P2 15; cf. P2 12, 16; 8.24). A ‘word’ ‘by itself can have meaning only at the level of spelling’ or ‘pronunciation’ (P2 91).

8.55 For these three motives, he falls back on an intermediary unit he ‘terms a “piece”‘, i.e., a ‘combination of words’ (P1 122, cf. P1 146, 192) (cf. Hallday's ‘group’, 9.75). His ‘main purpose is to guide’ not just ‘the descriptive analysis of languages’, but also ‘the synthesis for dealing with longer pieces of language’ (P2 202, 130; cf. 8.43, 64, 88). Such a ‘longer piece’ is a ‘meaningful complex’ to be ‘described as a relational network of structures and systems at clearly distinguished but congruent levels, converging again in renewal of connection with experience’ (P2 192; cf. P2 102, 122; 8.40). This recourse permits him to remain undecided about the ‘commonplace of linguistics’: that ‘the sentence and not the word is its main concern’ and ‘primary datum’ (P2 156, P1 170) (cf. 9¹³). He doesn't ‘define the sentence’, but does say what he thinks it is not. For him, ‘the sentence’ is neither ‘the lowest unit of language’ nor ‘a “self-contained” or “self-sufficient unit”‘ (P2 156); and his mention of ‘one-word sentences’ and ‘verbless sentences’ (P2 102) seems to depart from traditional or generativist definitions (cf. 4.67; 5.56, 58; 13.54). Nor will he accept a ‘definition of the sentence’ as a ‘“unit of predication”‘, a ‘judgment’, or a ‘proposition’, since he wants to ‘abandon’ ‘all this logical or psychological analysis’ (P1 170f, P2 102) (cf. 3.35f; 39; 4.69; 5.55; 8.17, 31; 9.14, 72; 11.39ff; 12.78f; 13.14).

8.56 Notwithstanding, he finds it ‘natural’ that ‘the sentence and syntactical analysis find a central place’ in a ‘general theory’ (P2 148). ‘When we speak’, we ‘use a whole sentence’, so ‘the unit of actual speech is the holophrase’ (TMS 83). Therefore, ‘the technique of syntax is concerned with the word process in the sentence’ (P1 183). He even approves Wegener's (1885) speculation that ‘all language elements are originally sentences’, such that ‘sentences’ might contain ‘the origins of all speech’, both ‘biographically’ and ‘historically’ (P2 148) (cf. 13.54). Yet this view conflicts with his usual search for origins in habits and social organization (cf. 8.6), and seems unrelated to such notions as the generativist ‘history of derivation’ (7.48), which is neither ‘biographical’ nor ‘historical’.

8.57 The indecision between word and sentence naturally carries over to Firth's views on organizing the levels of morphology and syntax (cf. 2.55, 61; 3.34f; 4.61ff; 5.53f; 6.49; 7³⁴; 9.31; 11.35; 12.77). At one point, he says they ‘have quite distinct sets of categories’, but later that ‘the distinction between morphology and syntax is perhaps no longer useful in descriptive linguistics’ (P1 6, P2 183; cf. 2.55; 3.34; 4.61; 7.58; 8.77; 13.28). In line with the second view, ‘morphological categories are to be treated syntagmatically and only appear in paradigms as terms or units related to elements of structure’ -- whence ‘the need for prosodic analysis’ (P2 183; cf. 8.64). ‘The study of syntactical categories’ should consider how ‘subordinate constituents are expressed and integrated’, and how ‘the utterance analyses or synthesizes the aspects of a complex situation’ in ‘the social conditions of employment’ (P1 223, 226; cf. TMS 190; 8.42).

8.58 Indecision also appears in his attempts to state how ‘the categories of morphology’ and ‘the parts of speech’ ‘arise from the formal conditions of the language’ (P1 24). He cautiously suggests ‘using fourteen parts of speech in a common grammar of careful polite English that can be written in orthography’ (P2 12), but doesn't tell us what they are, or how many might be needed, say, for careless rude English that durst not be written. He admits (with Richard Temple) that since ‘the functions a word fulfils in any particular sentence can be indicated by its position with or without variation of form’, a ‘word’ can ‘belong to as many classes as there are functions which it can fulfil’ (P2 142) (cf. 3.16, 22, 24, 33; 4.47, 49; 7.63; 12.25, 27). An imposing task for description looms here, the more so if ‘classes’ are to ‘include as complete a list as possible of examples’ as well as ‘indications of productivity’ (P1 223).³¹

8.59 In addition, the ‘verbal characteristics in the sentence’ ‘are rarely in parallel with’ the ‘verbal paradigms’ and ‘tabulated conjugations in the grammar books’ (P2 103f) (cf. 9¹¹; 13.27). In dealing with ‘operators’ (cf. 8.64), for instance, ‘tabulated paradigms’ like that of ‘the English verb’ ‘will get you nowhere’ (P2 121). He contemplates ‘putting aside’ ‘the paradigmatic approach to the morphology of separate words’, e.g., in ‘hyphenated lists of orthographic forms of individual words’, lest we ‘obscure the analysis of elements of structure in the syntagmatic interrelations of grammatical categories’ (P2 189, 203) (cf. 7.75f). Still, he allows that ‘systems’ of ‘units or parts of speech’ constitute ‘paradigms’ wherein ‘values’ are ‘given’ by ‘relations’ (P2 112, 173). He terms a ‘paradigm’ a ‘formal scatter’, ‘involving all the morphology of word-bases, stems, affixes, and compounds’ (P1 25, 15, P2 103, 122) (cf. 4.57f). These ‘scatters’ ‘can be arrived at’ ‘by recollecting, or by asking the native speaker, or by collecting verbal contexts’ (P1 25).

8.60 Unlike Hjelmslev and Chomsky, Firth denounces ‘general or universal grammar’, on the grounds that ‘grammatical meanings are determined by their interrelations in systems set up for that language’ and that ‘grammatical forms of a language are never in a strict sense parallel in another’ (P1 191n, P2 191, 65, 123; cf. 6.10; 7.19f; 8.19, 86; 13.18). ‘The use of traditional grammatical terms’ for ‘a wide variety of languages’ entails ‘the dangers and pitfalls’ of ‘personifying categories as universal entities’ and practising ‘bogus philosophizing in linguistics’ (P2 190; cf. 4.72; 8.5). ‘Every analysis of a particular “language” must determine the values of the ad hoc categories to which traditional names are given’.

8.61 Firth's own plan for ‘a new approach to grammar’ to match his ‘new approach to meaning’ turns out to be another ‘dispersal’ of ‘inquiry’ and ‘statement at a series of levels’ (P2 114, 117; cf. 8.49f, 52, 84). Accordingly, ‘the description of grammatical systems’ entails ‘graphematic, phonological, ‘morphological, and syntactic’ ‘criteria’ or ‘functions’ (P1 222f, 24, P2 118, 174). Instead of ‘grammatical analysis which deals with relations of the individual words’, our study should ‘look for verbal characteristics in the sentence as a whole’; after all, ‘the categories of grammar are abstractions from texts, from pieces or stretches of discourse’ (P2 188, 113, 121; cf. P2 183; 8.54; 12.71; 13.39). Regarding ‘criteria’ for ‘nominal and verbal categories’, for example, we ‘usually find that verbal features are distributed over a good deal of the sentence’ (P2 187). A favoured example is: ‘“he kept popping in and out of the office”‘, which is ‘grammatically close knit as a verbal piece’; it's hard to say ‘where's the verb’ (P2 103f, P2 121, 188). Similarly, ‘such categories as voice, mood’, ‘tense, aspect, gender, number, person, and case, if found applicable’ to ‘any given language’, should be ‘abstracted from, and referred back to, the sentence as a whole’ (P2 190). We would then not be perplexed when the ‘exponents’ are not ‘words or even affixes’, or are ‘discontinuous and cumulate’, as in ‘Latin’, ‘Swahili’, and ‘Modern Hindi’ (what ‘is traditionally referred to as “concord” or “agreement”‘ (P2 182, 190) (cf. 4.67; 9²⁰).

8.62 Despite the diversity implied, the ‘dispersal’ of ‘grammatical description’ is given a major restriction running throughout linguistics: it should ‘recognize only those linguistic distinctions which are formally expressed’ (P1 222) (13.54). By implication, meanings not distinguishable by form would get left out rather than ‘dispersed’ to other levels, and Firth is evasive on this point. He attacks Bloomfield's suggestion that ‘the study of meaning is the study of grammar’, yet says later that ‘grammar’ is ‘a study of meaning in generalized terms’ (P1 15, P2 118) (cf. 4.45ff, 50, 64). Similarly, the ‘grammatical’ ‘level’ is named in some lists of the ‘levels’ at which ‘statements of meaning’ are to be ‘dispersed’ (P1 19, P2 82, 92, 112, 124), but in others ‘morphological and syntactic’ appear instead (P1 192, P2 33, 118, 200f).

8.63 For Firth, the converse of ‘formal’ ‘criteria’ is ‘notional’ ones (P2 223). He wants to ‘put aside all notional explanations’ as a ‘manifest disadvantage’ fostering a ‘confusion of grammatical and semantic thinking’ which ‘clouds the precise statement of fact’ in ‘linguistic analysis’ (P2 189, 204, 152). They should therefore be ‘rigidly excluded’ from ‘grammatical or phonological’ ‘levels’ (P2 177). But some evasion occurs here too. ‘Notional elements’ might be ‘unavoidable for ‘classifying contexts of situation and describing’ them ‘as wholes’; here, ‘notional terms are permissible’, provided they do not ‘involve the description of mental processes or meaning in the thoughts of participants’, nor of ‘intention, purport, or purpose’ (P2 177f, 200; cf. 8.24, 45).

8.64 In between formal and notional, Firth introduces the term ‘colligation’ for a ‘syntagmatic relation’ and ‘mutual expectancy’ among ‘elements’ of ‘grammatical’ ‘structure’ (P2 186, 111, 183). For instance, ‘contemporary English’ has a category of ‘syntactical operators’ (like ‘“was, were, have, has”‘, ‘“do, does”‘, etc.) which ‘function in negation’ and ‘interrogation’; ‘all negative finite verbs are colligated with one of the operators’ (P2 182) (cf. 5.48; 9.79). Here again, ‘syntactic analysis must generalize beyond the level of the word isolate’ (cf. 8.54). And since the ‘relations of the grammatical categories in colligation’ do not ‘necessarily have phonological shape’, we are once more discouraged from ‘segmental analysis of the phonemic type’ (P2 182f) (cf. 8.31f, 53, 65).

8.65 ‘The investigation of words, pieces, and longer stretches of text leads to the prosodic approach’, which ‘emphasizes synthesis’ and ‘refers’ ‘features’ ‘to the structure taken as a whole’ (as ‘Sweet foresaw’) (P1 xi, 138; P2 100, 193). Firth hopes this approach can deal with a range of issues, including ‘syllable structure’,³² ‘stress’, ‘intonation’, ‘quantity’, and ‘grammatical correlations’, in ‘the piece, phrase, clause, and sentence’ (P2 193, 122, 100, 102, P1 130, 134). The approach is also ‘more comprehensive than traditional’ ‘word studies’, fits the ‘view that syntax is the dominant discipline in grammar’, and may be ‘useful grammatically’ and ‘practically in teaching pronunciation’, as well as relevant to ‘fieldwork on unwritten languages’, ‘the study of literature’, ‘literary criticism’, and ‘stylistics’, maybe even to ‘historic linguistics’ (e.g. for issues like ‘Ablaut’ and ‘laryngeals’) (P2 126, P1 137f; P2 101, 195) (cf. 8.6, 81, 83f). Moreover, by ‘regarding the elements of structure as prosodically interdependent and mutually determined’, this ‘approach’ is a clear alternative to ‘the American procedure by segmentation or succession’ of ‘units’, and can treat items which are ‘prosodically one’ (e.g., contractions like ‘“won't”’) or ‘holophrastic’ (e.g., ‘verbal pieces’) (P2 102, 104f, 193) (cf. 8.56).

8.66 To ‘state’ ‘prosodic features’ and ‘isolate prosodic groups’, Firth particularly commends ‘the technical resources of phonetics, both descriptive and instrumental’, plus those of ‘phonological analysis’ (P2 193, 100) (cf. 8.70). His reasoning is that ‘the systematic study of sounds’ is far more advanced ‘in modern linguistics’ than that of ‘prosodies’ (P1 123). In exchange, ‘phonological’ ‘analysis’, such as for the ‘features’ of ‘positions and junctions,’ can ‘more profitably proceed’ via ‘prosodies’ of ‘words, pieces, or sentences’ (P1 xi, 123, P2 100). By this route, Firth joins the other theorists who based their conceptions on the domain of sounds and articulation (cf. 2.17, 67, 70f; 3.14, 18; 4.29, 34; 5.42;, 5¹²; 13.26). He too finds the physical base reassuring: ‘the sounds and prosodies of speech are deeply embedded in organic processes in the human body’ (P2 90).³³ A patriotic motive is involved as well: ‘the English have excelled in phonetics’ (P2 137, 60, P1 120, 92) (8.15).

8.67 And so Firth salutes the study of sounds. ‘Every important advance during the last century’ was made by ‘attacking the problem from the phonetic side’ (P1 74) (4.79). His hero Sweet is called to witness that ‘phonetics’, ‘useless by itself’, is ‘the foundation of all study of language, whether theoretical or practical’ (P1 119). And ‘phonetic analysis has made possible a grammar of spoken language’, ‘in the face of’ which ‘classical grammar recedes into obsolescence’ (P1 23, TMS 136) (cf. 9.24; 13.33). In sum, ‘phonetics is one of the most practical of the social sciences’, providing ‘techniques for the study of utterance’, ‘systemic analysis’, ‘statement of linguistic facts’, and ‘establishment of valid texts’ (TMS 203, P1 145). So ‘morphology’, ‘syntax’, ‘descriptive grammar’, and ‘descriptive semantics’ must all ‘rest on reliable phonetic and intonational forms’ (P1 18, 3). Moreover, Firth's group has ‘developed general linguistic theory in close application to particular descriptions’ which ‘have in the main, been phonological’ (P2 126). He feels that ‘adequacy in the higher levels of linguistic analysis’ demands ‘the same rigorous control of formal categories’ as ‘in all phonological analysis’; and that ‘phonemic description should serve primarily as a basis for the statement of grammatical and lexical facts’ (P2 191; P1 222) (but cf. 7.46). It therefore seems harsh to fault ‘American structural linguistics’ for creating ‘a surfeit of phonemics’, and for ‘attempting’ to ‘directly develop’ ‘the analysis of discourse -- of the paragraph and the sentence’ -- ‘from phonemic procedures’ or ‘by analogy to such procedures’ (P2 160, 191; cf. 5.44; 8.31, 60).

8.68 By association, the multi-level scope of prosody naturally carries over into sound study. ‘Phonology states the phonematic and prosodic processes within word and sentence’ and ‘the phonetician links all this with the processes and features of utterance’ (P1 183, P2 175).³⁴ The ‘differential values’ ‘represented’ in ‘phonetic notation’ ‘may have ‘morphological, syntactical, or lexical function’; and ‘the identification and contextualization of the phonemes’ is ‘important’ for ‘studying’ ‘forms in morphology and syntax’ (P1 3f) (cf. 4.34f, 40; 5.35, 45; 12.81). ‘Grammatical classification limits and groups the data in parallel with phonological analysis’: ‘the exponents of some phonological categories may serve also for syntactical categories’, and those of ‘grammatical categories’ may ‘require’ ‘phonetic description’ or ‘notation’ (P2 192, 184, P1 3). ‘The interrelations of the grammatical categories stated as colligations form the unifying framework, and the phonological categories are limited by the grammatical status of the structures’ (P2 193).

8.69 If ‘phonological statements’ can profit from knowing ‘the grammatical meaning of the materials’, Firth must oppose Daniel Jones (praised for ‘carrying’ the work of ‘the English school of phonetics’ ‘to all parts of the world’) by maintaining that ‘phonetics and phonology must be linked with studies of meaning’ (P2 192, P1 166, P2 72; cf. P2 48, 86, 192, 14, P1 226n; 6.43). With the usual Firthian twist, he says ‘sounds direct and control’, but ‘do not hold or convey meaning’; ‘in the normal contexts of everyday life, the sounds of speech are a function of social situations’, from which ‘meaning is largely gathered by’ applying an ‘assumed common background of bodily habits’ (TMS 171) (cf. 8.21f).

8.70 For further ‘clarity of statement’ in sound study, Firth presents three ‘separate’ ‘terms: phonic, phonetic, and phonological’ (P2 99) (cf. 2.68; 4.30; 6.43; 8.70; 12.80).³⁵ ‘The phonic material’ comes from ‘the raw material of experience’ ‘in all its fullness’. A ‘description’ is then made ‘in the technical’ ‘language of phonetics’; ‘beyond this again’, ‘the phonological analysis’ ‘selectively’ works at ‘a different level of abstraction’, though its ‘categories, features, or units will have exponents describable in the phonetic language of description’ (P2 99; cf. P1 3f). So ‘phonetics and phonology’, though ‘differing in scientific levels’ and ‘systems of discourse’, ‘must work in harness’ (P1 145). ‘Phonetics’ is ‘the most specialized linguistic technique, tending to be ‘narrower and more abstract’, while ‘phonology’ ‘might be called ‘systemic phonetics’, ‘giving each sound a place in the whole phonetic structure of system’ (P1 34f). To prove he is ‘not a phonemicist and does not set up unit segments’ each ‘occupied by a phoneme’, Firth postulates whole ‘systems applicable’ to ‘elements of structure’ (P2 99).³⁶ Due to ‘the “context” of the whole’ ‘system’, ‘in actual speech situations the elements’ reflect ‘general attributes or correlation of articulation, such as length, tone, stress, tensity, voice’, as well as ‘styles’ (‘rapid, colloquial’, ‘familiar’, etc.) (P1 21, 42; cf. 8.44).

8.71 The question is then how to find a ‘scientific notation’ for the ‘transcription’ of ‘sounds’ (P1 109-13, TMS 150). ‘Alphabetic transcription of speech’ is ‘a highly abstract proceeding’ and cannot produce ‘an exact record of every detail of sound, stress, or intonation’ (P1 53, 3) (cf. 9.52). ‘The units of a phonetic transcription are best abstractions from utterances’ and may call for ‘terms and notations not based on orthography’ or ‘any scheme of segmental letters’ (P1 149f, P2 190). Indeed, ‘letters’ may ‘lead phonetics astray’ by not ‘corresponding with the facts of speech’ or by hiding ‘the overlapping and mutual interpenetration’ of ‘sounds, and the integration of movement for the whole word or phrase’ (P1 148, TMS 29, 39). ‘Separate letters’ can also foster the ‘hypostatization of the symbols and their successive arrangement’ in the ‘theories’ of both ‘historical and descriptive linguistics’, witness ‘the apotheosis of the sound-letter in the phoneme’ (P1 147, 125f, 123; cf. P1 21f, 71ff, 165; 2.69; 4.38, 45; 6⁸; 13.26). ‘Similarity of sound being no safe guide to functional identity’, Firth ‘abandons’ ‘the principle of “one symbol, one sound”‘, but all he can propose is ‘a store of good letters’ in ‘different founts of type’ (P1 51, 4f, 148, 146).

8.72 Despite all this concern for sounds, Firth asserts that ‘scientific priority cannot be given to spoken language as against written language’ (P2 30; cf. 2.21; 4.37ff; 6.50; 9.42f; 12.83; 13.33). As Trench (1855) remarked, ‘a word exists as truly for the eye as the ear’ (P2 90, P1 9n). Just as ‘all forms of written language’ have ‘the implication of utterance’, ‘all forms of speech have also the implication of writing for linguistic statement’ (P2 30f; cf. 8.44). ‘In a sense, written words are more real than speech’ in being ‘portable, tangible’, ‘material, permanent, and universal’ (TMS 40, 146). Though ‘written language’ does entail ‘an abstraction from insistent surroundings’ and its ‘context is entirely verbal’, it is still ‘immersed in the immediacy of social intercourse’ and ‘largely “affective”‘, and ‘refers to an assumed common background of experience’ (P2 14, TMS 174f) (cf. 8.47). In any ‘symbol’ like a ‘written form’, ‘the general and particular meet’, and ‘a high standard of literacy is the foundation of modern civilized society’ (TMS 30, 40, 135).

8.73 So ‘the actual forms of writing or spelling are a near concern for the linguist in dealing with his material’ (P2 31). ‘Orthography’ can ‘transcend the vagaries of individual utterance’, being ‘grammatically and semantically representative’ as well as ‘phonetically’ (TMS 48). ‘Grammar must concern itself with letters and marks’, because ‘spelling and writing’ present ‘the first level of structural analysis in sorting out the grammatical meanings of texts’ (P2 116; cf. 8.53). Also, ‘explorations in sociological linguistics’ use ‘the pedestrian techniques of the ABC as the principal means of linguistic description’ (P1 75). But Firth admits ‘the linguistic “economies” of speech are not those of writing’, and ‘it is impossible to represent fully to the eye what is meant for the ear’ (TMS 174, 146). ‘For the masses of people, too, the written language shows very little correlation with speech behaviour’ (TMS 116) (2.22).³⁷ ‘Spoken and written languages are two distinct sets of habits’: ‘ear language is intimate, social, local’, ‘eye language is general and nowadays everybody's property’ (TMS 198). Thus, ‘unwritten languages have a freedom of progressive economy’ (TMS 174f).

8.74 Special ‘study’ should be devoted to ‘world systems of writing’, such as ‘Roman’, ‘Arabic’, ‘Indian’, and ‘Chinese’ (P2 31). Sometimes Firth favours ‘the Roman alphabet’, calling it ‘the best’ ‘of all ABCs’, and urging its ‘universal adoption’ (TMS 136, P1 75).³⁸ It has ‘worked well from the days of a greater Rome to the present’, when ‘Western civilization is become world civilization’ (P2 68). It also has ‘merits as the framework for scientific linguistic notation’; it ‘lends itself to analysis and synthesis’, ‘produces easily recognized differentiated word-forms’, and ‘uses a comparatively small number of signs’ that can ‘suit the phonology and morphology of almost any language’ (P1 69). Yet at other times, he says our ‘alphabetic notation’ ‘does not rest mainly on modern acoustic and physiological categories, but on fictions’ ‘set up by grammatical theory’ (P1 148). And he contemplates ‘how much was lost’ by imposing ‘a theory of the Roman alphabet’ on languages in India, whose ‘syllabaries’ for ‘Sanskritic dialects’ are ‘models of phonetic and phonological excellence’ (P1 124f).

8.75 Moreover, ‘English’ ‘spelling’, though it may have ‘the longest literary tradition in Western Europe’, is a ‘handicap’, ‘preposterous’ and ‘disgraceful’ (P2 137, TMS 136, P1 112, 125). It ‘should be reformed’ ‘in the interests of the whole world’ (TMS 136, 48; P1 73), but Firth can't decide how. He is shy about adding ‘new letters’ or ‘written signs’ for fear of ‘swamping the characteristics of the alphabet’ or creating a ‘pepperbox spelling’ with ‘“accents”‘ and the like (P1 70, 124). ‘Purely phonetic spelling’ ‘is out of the question’, because ‘removing phonetic ambiguity’ ‘creates other functional (grammatical and semantic) ambiguities’ (P1 5, TMS 47f, P1 25) (cf. 2.69). And he shows no sympathy for ‘spelling pronunciations’, though they are ‘increasingly common’ (TMS 198f; cf. 2.21; 4.38).

8.76 Firth's diffuse schemes for a ‘language of description’, which I have essayed to review so far (8.35-75), come into sharper perspective in his advice about the ‘language under description’ (cf. 8.35). He vows that ‘descriptive linguistics’ ‘is at its best when applied to a restricted language’, which he defines as ‘serving a circumscribed field of experience or action’ and having ‘its own grammar and dictionary’ (or ‘a micro-grammar and a micro-glossary’) (P2 124, 87, 98, 105f, 112, i.r.). Such a domain is easier than ‘when the linguist’ must draw ‘abstractions’ from ‘a whole linguistic universe’ comprising ‘many specialized languages’ and ‘different styles’ (P2 30, 97, 118). ‘The material is clearly defined: the linguist knows what is on his agenda’, and can ‘set up ad hoc structures and systems’ for ‘the field of application’ (P2 106, 116; cf. 8.32). Once ‘the statement of structures and systems provides’ ‘the anatomy and physiology of the texts’, it is ‘unnecessary’ ‘to attempt a structural and systemic account of a language as a whole’ (P2 200).

8.77 ‘Linguistics’ can regard each ‘person’ ‘as being in command of a constellation of restricted languages, satellite languages’, ‘governed’ by ‘the general language of the community’ (P2 207f). As domains of ‘restricted languages’, Firth looks to ‘science, technology, politics, commerce’, ‘industry’, ‘sport’, ‘mathematics’, and ‘meteorology’, or to ‘a particular form or genre’, or to a ‘type of work associated with a single author or a type of speech function with its appropriate style’ or ‘tempo’ (P2 106, 98, 112, 118f) (cf. 9.106, 9⁴⁸).³⁹ By ‘promoting such restricted languages’, we may ‘advance international European cooperation’ and ‘unity’, e.g., among ‘teachers’ and ‘colleagues in various professions’ (P2 106). Ironically, Firth's own ‘successful application’ of ‘operational linguistics’ was not for unity, but for ‘air-war Japanese’ ‘during the Second World War’, when he assisted ‘the Royal Air Force’ (P2 29, P1 95, 125, 182) -- a motive perhaps in setting up the London chair (and one Mrs. Thatcher would have saluted) (cf. 8.1).⁴⁰

8.78 The ‘restricted language’ is a prime domain for discovering ‘collocations’: for ‘studying key words, pivotal words, leading words, by presenting them in the company they usually keep’ (P2 106ff, 113, 182) (cf. 9.93). This ‘study’ ‘may be classified into general or usual collocations and more restricted technical or personal’ ones, or into ‘normal’ and ‘idiosyncratic’ ones (P1 195; P2 18). ‘Characteristic distributions in collocability’ can constitute ‘a level of meaning in describing the English’ of a ‘social group or even one person’ (P2 195). Ominously, Firth's favourite demonstration word seems to be ‘“ass”‘, said of a person (‘collocation’ with ‘“silly”, “obstinate”‘, etc.) (P1 195; P2 108, 113, 150, 179). (Perhaps he should have lived to see the Thatcher government after all.)

8.79 ‘The collocations presented should usually be complete sentences’, or, in ‘conversation’, ‘extended to the utterances of preceding and following speakers’ (P2 107).⁴¹ Unlike ‘colligations’, ‘collocations’ obtain ‘between words as such’, not between ‘categories’ (P2 181, 69; cf. 8.64). Nor is ‘a colligation’ to be ‘interpreted as an abstraction in parallel with a collocation of exemplifying words in a text’ (P2 182f). But ‘the study of the collocation’ can be ‘completed by a statement of the interrelations of the syntactical categories within’ it (P2 23; cf. P1 xi).

8.80 ‘The study of the usual collocations’ ‘ensures that the isolate word or piece’ ‘is attested in established texts’, and provides ‘a precisely stated contribution’ to ‘the spectrum of descriptive linguistics’ by ‘circumscribing the field for further research’, e.g., by ‘indicating problems in grammar’ or aiding ‘descriptive lexicography’ with ‘citations’ for ‘dictionary definitions’ (P2 195, 180f, 196; cf. 8.43, 48). We should state ‘first the structure of appropriate contexts of situation’, ‘then the syntactical structure of the texts’, and ‘then’ ‘the criteria of distribution and collocation’ (P2 19). For example, ‘grammatical collocation and distribution provide differentiating criteria’ that ‘establish’ ‘the categories of noun substantive and verb’, or ‘guarantee the binary opposition of singular and plural’.

8.81 Firth recommends making ‘an exhaustive collection of collocations’ in ‘a restricted language for which there are restricted texts’ (P2 181). His own most ambitious attempt is far from exhaustive, however. He examines ‘English letters [i.e. epistles] of the eighteenth and early nineteenth centuries’ to contrast ‘collocations’ ‘recognized as current for at least two hundred years’ against those which ‘seem glaringly obsolete’ or ‘dated’ (P1 204f). This contrast is to be made by ‘applying the categories of context of situation and of meaning’, but his ‘outline sketch’ merely ‘suggests by hints’ how ‘a linguist’ might proceed ‘at a series of levels’ (P1 204, 214). For vocabulary, he turns to a dictionary of ‘Synonymes’ from 1824; for grammar, he focuses on the ‘“-ing” participle’, mainly ‘with preceding genitive’ (e.g. ‘“your trying it”’) (P1 205, 207-13). A handful of passages accompany the recommendation for a study of Dr. Johnson's ‘English in all his prose styles examined objectively and statistically’, plus ‘a biographical study of his personality’, to ‘give us a statement of stylistics in a social setting’ (P1 206).

8.82 Firth claims that he has ‘illustrated as many modes of meaning as possible from the language forms themselves’, and that ‘the name of a collocation is the hearing, reading, or saying of it’ (P1 214, P2 181). These claims suggest we need merely present the data to speak for itself (cf. 4.55). And such is clearly the method for his 1937 ‘revue’ of texts (from management, ‘technology’, ‘politics’, ‘business’, ‘advertising’, ‘religion’, etc.) (TMS 118-134). We are asked to ‘imagine’ the ‘response they would evoke’ in people past and present, including ‘feelings, social attitudes, prejudices, fears, fantasies, ambitions’ (TMS 117f), but Firth tells us little about his own response. Perhaps he feels he could not produce an analysis meeting his own demands for stringency, especially for complex and subtle materials.

8.83 Similar limitations beset Firth's sallies into ‘stylistics’, which he also ‘claims’ as a ‘level of linguistic analysis’ (P2 106; cf. 8.50, 64). Deriding ‘“discourse analysis”‘ ‘in America’, he wants a ‘much more systematic’ and ‘disciplined approach to the study of language’ and ‘literature’, stating ‘features or elements’ ‘in linguistic terms’ and ‘avoiding value judgment’ and ‘aesthetic appreciation’ (P2 106, 125; cf. P1 190, 202). ‘Style’ results from ‘fusing’ ‘elements of habit, custom, tradition,’ and ‘innovation’ within ‘verbal creation (P1 184) (cf. 3.69; 5.82; 6.52; 9.102; 11.57). Even if ‘whenever a man speaks’, it is ‘in some sense as a poet’, ‘poetry’ stands apart as ‘any piece of prose for which another’ ‘cannot be adequately substituted’ (‘attributed to Paul Valery’) (P1 193, P2 18, 25). ‘The poet so shapes his composition’ that ‘a great deal of its meaning is the form he gives it’ (P1 214) (cf. 3.69).

8.84 Not surprisingly, Firth disperses ‘stylistic analysis’ to a number of ‘levels’ and favours the ‘lower’ ones: ‘phonetic, phonological, prosodic, and grammatical’ (P2 195, P1 198, 200; cf. 8.51; 9.103). In his treatment of Swinburne, he recommends, but does not demonstrate, ‘starting’ from ‘the contextual study of the whole poem’ ‘by the methods of linguistics’ (P1 201). ‘Criticism at higher levels’ would attend not just to ‘the culture context’, including ‘biography and history’, but also to ‘word-formation or descriptive etymology’, ‘syntax’, and ‘phrasal stylistics’, e.g., ‘the association of synonyms, antonyms, contraries, and complementary couples’, ‘reversed and crossed antitheses’, or ‘patterns of opposition’ (P1 199-202). Since, however, ‘a detailed study of the words and pieces’ ‘would be laborious’, ‘scholars might be satisfied’ to ‘guess the probable result’, and Firth is content to do ‘without reference to the higher levels’ (P1 201, 203). He proffers the alibi that Swinburne is ‘the most “phonetic” of all English poets’ (P1 197, TMS 188).⁴²

8.85 Having covered the first two ‘methods of stating linguistic facts’ (‘language of description’ and ‘language under description’), we now come to the ‘language of translation’ (P2 49, 87, 98, 112, 149, 158, 202) (8.35). In its usual sense, ‘translation’ is a ‘science and art’ offering ‘a world-wide range for experiment’ and ‘inter-cultural co-operation’ (P2 135). ‘The fact of translation’ is both ‘a necessity’ on ‘general human grounds’ and ‘a main challenge’ to ‘linguistic theory’ to apply its ‘technical’ ‘description’ (P2 77, 66, 82; cf. P2 83, 197). But ‘in the widest terms’, ‘we are really translating’ ‘whenever we enter into the speech of someone else or our own past speech’, so we must account for ‘“translation” within the same language’ (P2 77f, 198).

8.86 ‘No translation is ever final or complete’ or ‘really equivalent’ (P2 79, 197, 76, 112). ‘One can never expect the modes of meaning’ to be ‘parallel or equivalent’ between ‘languages’, especially the ‘phonetic and phonaesthetic’ (i.e. sound symbolism),⁴³ and ‘universal grammar’ is of course rejected (P2 92, 196, 82; cf. 8.19, 60). But ‘translation problems’ can be solved ‘in the mutual assimilation of the languages in similar contexts of situation’ and in ‘common human experience’ (P2 87, 76, 82).

8.87 While ‘the basis for any total translation’ ‘must be found in linguistic analysis’, ‘the reverse process of using a translation as a basis for linguistic analysis at any level’ is an ‘error’ (P2 76, 157). Firth scolds ‘linguists who constantly make use of translation in linguistic analysis’ ‘without a systematic statement of the nature and function of the translation methods used’ (P2 83). Again, he singles out ‘Americans’: ‘ethnographic linguists’ (e.g. Voegelin, Yegerlehner, & Robinett 1954) who ‘confuse’ ‘translation with grammatical and collocational statements’; and ‘structuralists’ (e.g. Harris 1951) in whose work the ‘loose, impressionistic’, ‘casual’, ‘slipshod, and uncritical’ ‘use of translation vitiates linguistic analysis’ (P2 165, 197, 49, 204nf).⁴⁴

8.88 Firth prefers ‘statement of meaning by various forms of translation and definition’ (Malinowski 1935) (P2 198). As possible forms, Firth lists ‘bit-for-bit translation’, ‘interlinear word-for-word translation (sometimes described as “literal” or “verbal” translation’), and ‘free’ (or ‘running’ or ‘idiomatic’) translation’ (P2 149, 198).⁴⁵ A ‘“comparison”‘ of these can make ‘the text become quite clear’, aided by ‘“the contextual specification of meaning”‘ and ‘detailed commentary’, e.g. ‘phonetic and grammatical notes’ (P2 165, 149). ‘Translation’ serves here merely to furnish ‘identification names of language for isolates’, i.e., ‘reference labels’ (P2 197, 158, 33).

8.89 In tribute to his own spirit, I shall end my survey on a bright note. (I would be happy if I have covered just a fourth of Firth). ‘Again and again, linguistic scholarship has served the formal education of its time’ (P2 130), and he hopes it will again (cf. 13.64). ‘Linguistic methods’ can ‘help the study of the Mother tongue’, combat ‘speech defects’, or develop ‘orthographies for Oriental and African languages’ (P1 93). Closely related tasks include ‘translating into Asian and African languages for education’ and ‘specialized occupations’, plus ‘collecting and collating traditional oral literature and other creative compositions’ (P2 135). And if ‘every cultured man needs a second and perhaps a third foreign language’, linguistic methods can ‘help learners’ to ‘acquire good pronunciation’ or to learn ‘a foreign language for reading only’ (TMS 211, P1 93, TMS 136) (cf. 4.86; 9.1).⁴⁶

8.90 Firth sees ‘the tasks’ of ‘linguistics’ ‘increasing in responsibility’ (P2 132). At such a time, ‘the greatest need of linguistic scholarship’ is ‘a new outlook over a much wider field of life’, and ‘new values’ (P1 32, 29f). We should ‘try other theories’, ‘overhaul our descriptive instruments’ and ‘languages of description’, and seek ‘more accurately determined linguistic categories for principal types’ of ‘usage in various social roles’ (P2 189, P1 28). ‘Vast research’ also awaits in ‘the biographical study of speech’ and in ‘sociological linguistics’ (P1 29, 27).

8.91 Above all, ‘general linguistic theory’ should be made useful in ‘describing particular languages and in dealing with specific language problems’ (P2 130). It should ‘guide’ ‘analyses’ and ‘provide principles’ for ‘synthesizing’ ‘the useful results of linguistic studies of the past’ (P2 130) (cf. 1.6f; 13.64). It should undertake ‘a serial contextualization of our facts, context within context, each one a function’ ‘of a bigger context, and all contexts finding a place’ in ‘the context of culture’ (P1 32) (cf. 8.26, 42, 44; 9.22f; 13.62). And it should also ‘produce the main structural framework for the bridges between different languages and cultures’ (P2 130, 202). To meet such tasks, Firth calls for a ‘linguistically centred social analysis’; ‘a description of speech and language functions with reference to effective observable results’; and a ‘study of conversation’ to seek ‘the key to understanding what language really is and how it works’ (P2 177, 112, P1 32). Much remains to be achieved in the ‘study’ of ‘language’, ‘and we are still far from understanding how it functions’ (TMS 147).

NOTES ON FIRTH

¹ The key to Firth citations is: P1: Papers in Linguistics 1934-1951; P2: Selected Papers of J.R. Firth 1952-1959; and TMS: Tongues of Men and Speech (London: Oxford, 1964 [1937 and 1930]). The works often repeat passages or overlap.

² Strevens notes that Firth's early work differs from later in views about ‘grammar’, ‘spelling’, and ‘experimental phonetics’ (TMS viiif). Further indecisions or contradictions involve: whether linguistics should be a separate discipline (8.16f); whether to use old or new terms (8.37); whether levels are mutually prerequisite (8.51); whether grammar should be stated in terms of meaning (8.45, 62); whether the sentence is the basic unit (8.54f); whether morphology and syntax should be merged (8.56); whether the roman alphabet is good or needs reform (8.74f); and whether collocations concern syntactic form (8.78f). To smoothe the lines of argument, I consign some contradictory statements to footnotes (cf. Notes 7, 10, 18, 26, 30, 33, 39, 43).

³ Pike's quotes of Firth are: ‘schematic constructs have no ontological status’ (LB 56; P1 181); ‘each fact finds its place in a system of related statements, all arising from theory’ (LB 56; P2 43); and ‘even in mathematics the possibilities of complete axiomatization have been overestimated’ (LB 56; P2 44). Pike finds ‘Firth's ‘place and order’ similar to ‘etic slot’, though the emphasis on categories over words (8.53) suggests an ‘emic’ view; and he identifies Firth's ‘prosody’ with ‘phonemics’, which Firth denies (LB 420, P2 27f). In return, Firth mentions Pike's differentiating the ‘etic’ from the ‘emic’, and his ‘grammatical prerequisites of phonemic analysis’, but protests that ‘all levels are mutually requisite’ (8.51); and sees Pike's ‘procedural approach’ as ‘an