Published online by Cambridge University Press: 22 April 2026
Distinctive features occur in lumps or bundles, each one of which we call a phoneme. The speaker has been trained to make sound-producing movements in such a way that the phoneme-features will be present in the sound-waves, and he has been trained to respond only to these features and to ignore the rest of the gross acoustic mass that reaches his ears. Leonard Bloomfield (1933)
The number of different phonemes in a language is a small submultiple of the number of forms. Leonard Bloomfield (1926)
The logical demand that a science speak in quantitative terms is met by linguistics because it speaks in terms of phonemes. Leonard Bloomfield (1927)
1 See, in particular, C. E. Shannon and W. Weaver, The mathematical theory of communication (Urbana, 1949); D. M. Mackay, In search of basic symbols, Cybernetics: Transactions of the eighth Conference (New York, 1952) ; id., The nomenclature of information theory, ibid.
2 This analysis was made as part of the research on contemporary Russian conducted by the Department of Slavic Languages and Literatures at Harvard University under a grant from the Rockefeller Foundation. The first volume of the description of contemporary Russian based on this research, dealing with the various aspects of Russian speech sounds, is now being prepared for publication. Grateful acknowledgment of help is also made to the Signal Corps, Air Materiel Command, Office of Naval Research, and the sponsors of the Fulbright program.
3 A. Peškovskij, Desjat' tysjač zvukov russkogo jazyka, Sbornik statej 167-91 (Leningrad, 1952).
4 For further information on distinctive features and their acoustic and articulatory correlates, see R. Jakobson, C. G. M. Fant, and M. Halle, Preliminaries to speech analysis, 2d printing (MIT Acoustics Laboratory, Technical report No. 13, 1952).
5 Cf. A. Isačenko, Fonetika spisovnej ruštiny 177, 182 (Bratislava, 1947).
6 We follow the IPA system of transcription, except in three respects: we use a comma after a letter to indicate palatalization ; we place the accent mark immediately before the vowel letter; and we render the strident stop by the same letter as the corresponding constrictive with the addition of a circumflex.
7 There are two competing varieties of contemporary standard Russian. The more conservative is codified especially in Tolkovyj slovar' russkogo jazyka, ed. by D. Ušakov (Moscow, 1935-40); the other is advocated in particular by S. Obnorskij, and is presented in Slovar' russkogo jazyka, ed. by S. Ožegov (Moscow, 1949). In general we accept Usakov's norms; but in order to include all the phonemic discriminations possible in standard Russian, we add to his traditional repertory of phonemes a new phoneme /g,/ as distinguished from /g/. Such new gerund formations as /b,ir,ig,|a/ ‘taking care’, distinct from /b,ir,ig|a/‘banks’, are admitted into standard Russian by Obnorskij and his followers.
8 Among Russian compound words we include all words with a non-initial root : words with more than one root, e.g. /adna-abr|aznij/ ‘uniform’; words with prefixes, e.g. /za-astr,|it,/ ‘to sharpen’, /iz-vad,|it,/ ‘to exhaust’; and words with preceding prepositions which are phonemically treated like prefixes, e.g. /za-akn|o/ ‘behind the window’, /iz-vad|i/ ‘out of the water’.
9 Cf. Shannon and Weaver 102: 'A system which produces a sequence of symbols ... according to certain probabilities is called a stochastic process, and the special case of a stochastic process in which the probabilities depend on the previous events, is called a Markoff process or a Markoff chain.' In his Essai d'une recherche statistique sur le texte du roman 'Eugène Oněgin', illustrant la liaison des épreuves en chaîne, Bulletin de l'Académie Impériale des Sciences de St. Pétersbourg, Vol. 7 (1913), A. A. Markov studied the distribution of vowel and consonant letters in a part of Puškin's famous poem and showed that the transitional probabilities between the letters were not those of a random sequence but rather depended on the preceding letter or letters.
10 David A. Huffman, A method for the construction of minimum redundancy codes, Proceedings of the IRE 40:9.1098-101 (1952).
11 Markov, in his study of letter distributions in a Russian poem, obtained the value 0.4317 for vowels and 0.5683 for consonants. His figures are remarkably close to ours, especially if we make allowance for the fact that Markov counted some instances of /j/ as vowels and others not at all.
12 For a fuller discussion, see R. Jakobson, Sound and meaning (to appear).
13 In R. Carnap's terminology, the occurrences of phonemes, having been studied in the Russian word-events, are to be investigated in the word-designs, just as we have here studied the occurrences of distinctive features in the phoneme-designs; cf. Introduction to semantics 3 (Cambridge, Mass., 1946). Charles S. Peirce, the founder of modern semiotic, would say that besides the application of the phonemic legi signs within the lexical sinsigns, such an application must be scrutinized again within lexical legisigns; cf. his Collected papers 2.245-7 (Cambridge, Mass., 1932).
14 An exhaustive statistical analysis of the phonemic structure of Russian root morphemes is being prepared by Robert Abernathy within the framework of the research program mentioned in fn. 2.