This book "supplies a vocabulary of English words and idiomatic phrases 'arranged … according to the ideas which they express'. The thesaurus, continually expanded and updated, has always remained in print, but this reissued first edition shows the impressive breadth of Roget's own knowledge and interests."
AUTHOR: Jones, Daniel TITLE: The Phoneme SUBTITLE: Its Nature and Use PUBLISHER: Cambridge University Press YEAR: 2009/1967
Arnaud Fournet, La Garenne Colombes (France)
The book was first published in 1950 but the author indicates in the Preface that the manuscript was ready in 1946. The reviewed copy dated 2009 is an unmodified reprint of the 1967 third edition which differs from the first one (at least) because of an extra appendix dealing with a historical sketch of the term ''phoneme'' and a number of modifications, such as the addition of §46a mentioning a book published in 1958.
Although British, the author, Daniel Jones (1881-1967), is not a pupil of Henry Sweet, the well-known English phonetician, as might be expected, but one of the French phonetician Paul Passy. Among other activities, the author was an active member of the International Phonetic Association for more than half a century, of which he ultimately became the president for 17 years from 1950 to 1967.
The book was designed by the author as a kind of treatise describing his approach to phonetics and phonology and ''setting out in some detail the theory of the phoneme and illustrating it with as many languages as [he] knew anything about'' (p.VI). This endeavor is reminiscent of Trubetzkoy's ''Principles of Phonology'' and a comparison between the two books, published in about the same decade, seems inevitable in a review.
The book is about phonetics and phonology. The approach of the author can be described according to two main criteria: it has the phoneme as its main descriptive tool and it is thoroughly empirical as regards methodology. Jones considers phonology to be a section of phonetics, not a separate field. It differs from ''(post)-generative'' approaches which rely on the (Jakobsonian) feature as the main descriptive concept and different types of (SPE-inspired) axiomatic formalism to describe phonemes, features and (morpho-)phonology. It can be added that the phoneme is defined by Jones as a segment, a clipping in the chain of speech sounds, which also makes his approach different from that of Bloomfield. Some phrases like ''we think we utter/hear'' (§4) or ''I consciously put [stress]'' (§453) are also very far from Bloomfield's approach and American structuralism. In addition the segmental approach of Jones implicitly rejects any notion of distinctive feature: ''minimal distinctions'' (Chapter VI) include segments, vowel length, tone, position of stress but do not include for example voice or aspiration. A fortiori Jones does not address issues of mark or markedness. In footnote 12 p.85 he states that a theory of the phoneme based on features ''would have little practical value even if it could be worked into a consistent system''. Jones has an approach to phonology extremely influenced by issues of alphabets and writing, which is probably due to his involvement in the IPA. Cf. p.253 and §27 p.264. In this framework the phoneme is a clipping which cannot be further analyzed. On the whole the book is densely packed with data and concrete discussions of phonetic phenomena and much less deals with issues of method and theory, which have to be inferred, whenever possible, from what Jones states. Most examples are taken from (British) English, (Parisian) French, and less frequently German, Italian, sometimes Chinese (Pekingese or Cantonese) and miscellaneous other languages. Tswana is often used when discussing prosodic issues.
The book is rather conventional, comprising a Preface, acknowledgments, contents, 32 chapters with a total of 769 numbered paragraphs, an appendix dealing with the term ''phoneme'', a bibliography, two indexes for languages and subjects. To a large extent each chapter can be read independently. The author sometimes refers to his other works, which is somewhat of a hindrance. The book is a continuous flow of chapters and paragraphs without any suggested segmentation. It is possible to distinguish four main parts. The first part (I to VI) is a theoretical introduction which delves into general considerations about language, speech and sounds. The second part (VII to XXV) deals with a large number of phonetic phenomena and examples. It could be further divided into two subsections dealing with ''segmental'' (VII to XX) and ''supra-segmental'' (XXI to XXV) issues, although these are not words used by Jones himself. The third part (XXVI) is typological. The fourth part (XXVII to XXXII) is about general considerations, especially of social and historical nature. The chapter (XXXI) which discusses phonetic writing also deals with issues like orthographic reform.
It is not possible to review all the data and discussions presented in the book. I will focus on the author's approach and its coherence, its avowed and actual relationships with contemporary approaches and the accuracy of some key assertions of the author. To begin with I tend to think that it would be highly desirable to publish an annotated edition of the original book, as was once done with Saussure's ''Cours de linguistique générale'' (CLG) between the 1916 and 1964 editions. On the whole I found the book interesting, entertaining, stimulating and at the same time disturbing. The reasons for these reactions will gradually appear below. There is a general problem throughout the book: most of the framework of the author is implicit and what is more, the real framework underlying many assertions and operations differs from the avowed framework.
Generally speaking phonology as described by Jones is a ''classification'' (p.42) of sounds into increasingly larger groupings and a phonological survey is to be understood as a classification and a transcription of sounds (§39, §666). At the bottom of the hierarchy of inclusion are actual ''speech-sounds'' or ''phones'': segments more or less adequately carved out in the ''speech chain'' (p.1). The issue is to ''assign sounds to phonemes'', or to determine ''which phoneme a sound belongs to'' (p.41), typical and oft-repeated phrases. Jones distinguishes two levels of sounds: ''concrete'' and ''abstract'' in a way that amounts to token and type (§20). A phoneme is defined as ''a family of sounds which are related in character'' (§31, §649, p.261). The relationship can be either ''acoustic or organic'' [= articulatory]. He emphasizes ''the importance of avoiding to introduce non phonetic criteria'' (p.64) in assigning sounds to phonemes, a point that needs to be commented on as shown below. §649 speaks of a ''physical'' approach to the phoneme. It could be added that ''phonetic'' for Jones means articulatory much more than acoustic, as the diagrams in the book always refer to positions of articulation and there is no mention of formants.
Jones seems to be aware that there is some intrinsic difficulty in his definition. For example he considers English [h] and velar nasal [N] too different ''in character'' to be the same phoneme, which seems fairly acceptable. At the same time he groups together English [t] and glottal stop [ʔ] as the same phoneme, in spite of their irreconcilable dissimilarity which he tries to downplay (p.10) and the fact that these sounds are different phonemes in many languages. Cf. §618 for a suggestion that [ʔ] could “belong to” three different phonemes in English. Contrary to what Jones asserts in §49 (very) different sounds can appear in the same phonetic context and nevertheless ''belong to'' the same phoneme, as seen above. This situation is a serious threat for the ultra-phonetic ''physical'' approach apparently advocated by Jones. This raises the issue of a phonological theory that would rely exclusively on phonetic factors to define phonemes. In all cases (in my opinion) Jones did not succeed in developing a theory on that premise in the book.
There is a problem of internal coherence or hidden circularity in the method followed by Jones. He considers semantics to be a ''function of the phonemes'' (chapter IV, §52), ''what they do'' (p.265), derived from their phonetic properties and not a key component of their definition. But the rationale which justifies the segmentation of the chain into speech sounds is semantics: ''the sections being such that [their] exchange is capable of changing a word into another word'' (p.3). Is a word not (at least) a phonetic form plus a meaning? Jones hardly ever mentions the terms commutation or substitution but in practice he applies the method of analysis typically used in European structuralism and the Saussurian-Prague school. Here we perceive a contradiction: Jones tends to portray himself as independent of Saussurian-Prague structuralism but in practice most of what he does or writes is similar or identical to it. §7 and §655 (and also §51) clearly expose the non-existence of a supposedly independent alternative approach of phonology portrayed by Jones. This contradiction, which is reflected by ''I came to the phoneme entirely through phonetics'' (p.264), appears in numerous sentences. For example: (1) ''it is constantly found in language study that several distinct sounds have to be considered as if they were one for orthographic, grammatical and semantic purposes.'' (§21); (2) ''whether a vowel heard as long is to be considered double is a matter to be decided by the feeling of the speaker, which in its turn is determined by reference to the structure of the language'' (§363); (3) ''the values attached to the letters in phonetic transcription are not absolute but must always be interpreted in the light of the phonetic context'' (§312). How can these statements be reconciled with the supposedly ''physical'' approach that Jones claims to follow? It is obvious that Jones follows a Saussurian-Prague approach much more faithfully than what he is ready to acknowledge. §613 even mentions sounds that are recognized ''not so much by their inherent quality as by their relationships to the speaker's other sounds'' in a completely Saussurian wording: compare ''to know the phonic units of a language it is not indispensable to characterize their positive quality; it just takes to consider them differential entities whose main feature is not to be confused with each other'' (CLG p.303).
A second pillar of the method proposed by Jones is the distinction between so-called ''narrow'' and ''broad transcription''. Narrow transcription can be defined as a detailed phonetic description (of a professional phonetician type) and broad as a more minimalist one (as is useful for teaching purposes, Cf. §664). The claim made by Jones is that his broad transcription is phonemic. But there are serious reasons to doubt it. For example Jones consistently writes French vowels as long when followed by a word-final -r#. This seems to be also the case in Passy's Dictionary of French phonetics. Vowel length in that context is not phonemic. This leads him to describe a ''shortening'' between 'faire' and 'faire part' supposed to be /fe:r/ and /fer pa:r/ (§386, 394). The conditioning factor for length is standard French accentuation on the last syllable.
Another problem with Jones' broad transcription is the issue of diphthongs or sequences. For example Jones considers English 'ay' to be a diphthong /ei/ but French 'eil' a sequence /ej/. But what about English 'to pay' versus French 'paye', which are not discussed by Jones? If the criteria are phonetic or ''physical'' as Jones claims, why should English 'pay' be /pei/ but French 'paye' be /pej/ when both sound and mean (nearly exactly) the same? Cf. §243, 244 and footnote 9 p.74 for this issue. Apart from the absence of explicit criteria to explain the difference in the book, this point also raises the general issue of the distance or abstraction from phonetic substance that is acceptable or necessary in a broad description. For example, Jones mentions the fact that there is no **[tu] or **[ti] but only [tsu] and [tʃi] in Japanese and criticizes the Hepburnian transcription of Japanese for introducing ''unnecessary complications in spelling'' (§.140).
Something missing from the book in connection with diphthongs and affricates is the syllable, which Jones never defines but indirectly and sketchily depicts as follows: ''We know how many syllables there are in a word by counting the peaks of prominence, but we cannot define with precision the points at which syllables begin and end.'' (§464) Jones does not propose any criteria as to why [ts] or [tʃ] should be considered an affricate or not in a given language. In justice to Jones it must be mentioned that Trubetzkoy likewise did not define what a syllable is in his ''Principles of Phonology''. It can nevertheless be observed that Jones does not place the stress mark (') at random and that there is an implicit approach of syllabic segmentation at work. §551 talks about ''possible Italian combinations'' of phonemes but does not develop the idea. In relationship with the issue of syllables and syllabic structure, Jones describes French 'droit' as /dRwa/ but I tend to think it is much more adequate for several reasons to deal with [wa] as being a diphthong /ua/. That is another point where his broad transcription remains phonetic when a phonological description or representation should or could be otherwise. Jones does not discuss this issue although he mentions that ''a vowel and a consonant occasionally occur as members of the same phoneme'' (§281), a statement that raises quite a number of practical and theoretical issues. The missing syllabic approach also leads to some oddities such as French trahi /trai/ and travail /tRavaj/ being used to oppose vocalic /i/ and consonantal /j/ (§285). In justice to Jones it must be mentioned that other well-known phonologists like Martinet have asserted similar things (with 'abbaye' /abei/ and 'abeille' /abej/ with respectively three and two syllables).
Another issue where Jones' approach finds considerable difficulty is the case of sounds which are phonetically different from all others and cannot be easily assigned to any particular phoneme on the basis of complementary distribution. A typical case discussed in §195 is the Chinese (Pekingese) sibilant written 'x' in Pinyin. To further compound the problem 'x' is one of the three sounds that can be followed by the (phonetic) vowel [y]. Jones proposes to assign 'x' to /ʃ/ which means an extra vocalic phoneme /y/ in addition to /u/ with a massively lacunary distribution. The situation is made yet worse by the existence of the affricates (Pinyin) 'j' and 'q' which Jones does not discuss. This issue is also connected with the concept of neutralization and archiphoneme, which Jones calls di-phonemic sounds (Cf. Chapter XX). This is basically Saussurian-Prague phonology worded differently. In addition it would be necessary to talk about tri- or quadri-phonemic sounds as in the canonical case of preconsonantal nasals in Dravidian but Jones does not address this example. Jones does not talk about neutralization and considers on the contrary that a sound should be preferably assigned to one phoneme even if that means ''arbitrary'' assignment (§338). Apart from the principle itself this approach undoubtedly raises a number of practical issues. For example the Russian phonemes /o/ and /a/ are distinguished only when stressed. Jones assigns Russian unaccented /o/ to /a/ which means ''a change of phoneme'' between for example the masculine xopoши [xa''roʃi] and the neuter xopoшo [xaraʃ''o] (Cf. §539). The same problem appears in French between the different varieties of /e/ (closed, open and (archaic) long) which are neutralized when not word- or syntagm-final (§221 and footnote 7). Jones does not discuss another canonical and frequent example of word-final devoiced consonants. (Cf. §332 for uncommented Russian examples).
Another issue which seriously threatens the avowedly purely ''physical'' approach of Jones is the interferences between phonology, morphemic structure and syntagmatic morphology. A typical case discussed in the book is Scottish English (§377, 522) which displays a number of phonetic peculiarities when '-es' and '-ed' are suffixed to a base. Another instance is French 'oui' and 'ouest' considered to be /wi/ and /west/ by Jones (§283). The issue here is that the initial of 'oui' blocks connection or elision with the articles 'un, le # oui' but 'ouest' accepts them: 'l'ouest'. The same problem exists with 'héro' and 'erreur' (§307). Jones does not even seem to see that there is an issue here: where is the right place to deal with these phenomena: phonology, morphology, somewhere else?
A number of paragraphs of the book also raise different issues: (1) how legitimate are syntagms or proper names in a phonemic analysis? (Cf. footnote 23 p.65); (2) on the basis of the pair of German words Elend ~ fehlend Jones proposes to postulate a schwa phoneme that would exist only in unaccented syllables (§227). This raises the issue of having more phonemes in unaccented than accented syllables. Jones does not seem to be uncomfortable with this consequence; (3) in §286 and §366 Jones disregards the difference between long and geminate consonants, postulating in §364 that the difference never plays any phonemic role and he cuts Italian long consonants in two parts assigning each half to a different syllable (§286); (4) §9-13 contain a number of phonetic definitions which are inadequate and contradictory ; (5) §490 is quite mysterious: ''If minimal distinctions were always effected by changes in a single attribute [=feature], there would be no phonemes.''
At this point of the review it could be assessed that from a sheer technical point of view Jones' treatise is interesting, rather richly documented but at the same time it fails to adequately address a number of canonical and recurrent issues and theoretical questions of phonology as a scientific field. And the next point is that the main problem with the book is its covert and implicit framework. And this problem is in fact a bundle of problems.
The first problem of epistemological nature is that the real framework is not explicit. On the surface it would seem that Jones did not really believe that a precise terminology was either possible or necessary: ''certain fundamental concepts in phonetics, as in other sciences, are incapable of precise definitions.'' (p.V). The approach developed by Jones may appear to be nearly anti-theoretical. The general framework and most key concepts are worded in everyday English. A first consequence of this situation is a pervading lack of rigor and opaque methodology. A second consequence is that the dialog between concrete data, methods and concepts is crippled. In addition §87 defends a ''mentalistic view of the phoneme'' in discrepancy with the rest of the book. §137 contemplates regarding the phoneme as an ideal ''sound to be aimed at'' but Jones seems unconvinced by the idea. Cf. 625 as well. Some paragraphs are contradictory: §333 is against etymology or derivation as a clue to phonemic analysis but §367 and §329 state the contrary; §217 emphasizes the idiolect for phonemic analysis for French but §331 states the contrary in the case of Japanese. Cf. §624 as well.
The second problem is that Jones claims to propose an alternative to Saussurian-Prague phonology (§649) but his implicit framework defines the phoneme as a segmental unit extracted in the chain by commutation, just as in Saussurian-Prague phonology (Cf. §7, §144 and §655). A conspicuous feature of the book is that Jones almost never used words which have belonged to the basic core of technical terms of linguistics for decades. Cf. §5 p.263 for feedback on these old-fashioned words used by Jones. The verb 'to articulate' is also rare, especially in a book dealing with phonetic issues. The distribution of a phoneme is called ''manner of use''. The term ''idiolect'' which has existed in English at least since 1800 is replaced by the ambiguous and misleading ''style of speech''. Toward the end of the book the terminology increasingly resorts to standard structuralist words: ''structure'' and ''system'' appear respectively on p.220 and p.229. On p.105 Jones talks about ''symmetrical use of the vowels'': what does this wording mean if Jones' approach is purely ''physical'' (§649) and is not about phonological systems and structures? It can also be noted that the concept of ''diaphone'' (chapter XVII) is impossible within the ''physical'' approach which Jones claims to follow. Similarly impossible in his ''physical'' approach is the notion of ''overlapping of phonemes'' (chapter XIX). Jones states that “it follows also from the physical definition that phonemes are of necessity units of linguistic structure.” (p.265). I see no logical connection here and Jones provides no explanation to substantiate this very strange claim.
Ultimately the absence of an explicit framework and the use of everyday English generate the impression of reading a learned but amateurish book written when linguistics was still in early infancy. The reviewed copy is a reprint of a book originally published slightly more than half a century ago, so it is more or less inevitable that reading or reviewing such a book entails some kind of travel through time. The contents and the terminology actually make the travel look like it is much longer than just half a century ago. In my opinion the book sounds like it was written well before Trubetzkoy's ''Principles of Phonology'' and not a few years after it. And it could be, to some extent, advised as a preliminary reading before tackling the real thing with Trubetzkoy's magnum opus.
The third problem is that the claim of an independent theory is not just a refusal to acknowledge intellectual debt. §656 is sarcasm aimed at Trubetzkoy himself, who is portrayed as a kind of clever but abstruse trickster and a marginal phonologist among others. From the book it would appear that the most towering figure of all the history of phonology so far, Roman Jakobson, never existed. His name and works are nearly nowhere mentioned or alluded to except on p.266 and footnote 14 p.257. The short footnote on Saussure on p.257 is also benchmark.
Now I come to the fourth problem. With the appendix added in 1967 Jones embarked on a new task with the third edition which nearly amounts to rewriting history. Most statements in that appendix conflict with facts for which there is conclusive and detailed evidence: (1) the person who coined the word ''phoneme'' was Dufriche-Desgenettes and not Havet; (2) Baudouin de Courtenay and Saussure met in Paris and had numerous discussions; (3) a continuous correspondence between Kruszewski, Baudouin de Courtenay and Saussure took place between the Mémoire written by Saussure in 1871 and the Cours published in 1916. On the contrary Jones talks about a supposedly large correspondence between Passy and Baudouin for which he mentions there is no record (p.256); (4) Passy followed Saussure's course in the EPHE; (5) Jakobson clearly stated well before WWII that the members of the Saussurian-Prague school originally did not know about Baudouin de Courtenay's approach; (6) Jones does not even mention the 1st International Congress of Linguists in The Hague in 1928 where Trubetzkoy, Jakobson and some others proposed and described the heuristic program of phonology in its modern sense. On the whole this part of the book is unpleasant and disturbing to read.
As a general conclusion, I must say that I have mixed feelings about the book. On the one hand, in spite of a number of oddities and contradictions the book could be recommended as a Primer to Saussurian-Prague Phonology written in everyday English without too much theory and without potentially obscure and ultra-technical jargon. But a serious problem is that it does not acknowledge its real conceptual, historical and epistemological roots: the Saussurian-Prague school. It is presented as an independent theory to the point of trying to rewrite history and to erase a number of major figures of linguistics and phonology such as Saussure, Trubetzkoy and Jakobson, to whom Jones is clearly and deeply indebted. In my opinion a consistent set of comments and annotations which signal the contradictions, warn against some inadequacies and neutralize some paragraphs and footnotes of the book would be highly useful.
De SAUSSURE Ferdinand. Cours de linguistique générale (CLG). Paris: Payot, 1965 (1916). English translation: Course in General Linguistics, Open Court Classics. Peru (Illinois): Open Court Publishing, 1986 (3rd ed).
TRUBETZKOY, Nikolai Sergeyevitch. Principes de phonologie. Paris: Klincksieck, 1970 [1st edition in German 1939, English translation: Principles of Phonology. Berkeley: University of California Press, 1969].
ABOUT THE REVIEWER
ABOUT THE REVIEWER:
Arnaud Fournet, unaffiliated scholar, La Garenne Colombes (France). His
research interests include (historical) phonology, descriptive linguistics
and macro-comparison, especially in the Nostratic perimeter.