¹ Cf. Leonard Bloomfield, Language 161, 209 f., 221 f. (New York, 1933); Otto Jespersen, The philosophy of grammar 97–107 (New York, 1924). Needless to say, my development owes a great deal to these sources.

² The first stimulus for this paper was received while I was attending the Linguistic Institute at the University of Michigan in the summer of 1945, holding a scholarship from the American Council of Learned Societies. An earlier version of the paper was read at a meeting of the Linguistic Society of America, 30 December 1947. I am indebted to Zellig S. Harris for help in preparing the published draft.

³ Cf. especially Bloomfield, op.cit.; Kenneth L. Pike, Taxemes and immediate constituents, Lang. 19.65–82 (1943); Pike, Analysis of a Mixteco text, IJAL 10.113–38 (1944), esp. 120 ff.; Eugene A. Nida, Syntax 44–61 (Glendale, Calif., 1946); Rulon S. Wells, Immediate constituents, Lang. 23.81–117 (1947). I am especially indebted to Pike and Nida for instruction in the principles of immediate constituents.

⁴ This useful abbreviation for immediate constituent (plural ICs) is taken from Wells, op.cit.

⁵ Of course there are always exceptions, for instance factors of intonation, concord, and substitution (pronouns). The last two might be handled by the method suggested in Zellig S. Harris, Discontinuous morphemes, Lang. 21.121–7 (1945), and the first by the method suggested in the same writer's Simultaneous components in phonology, Lang. 20.181–205 (1944).

⁶ On have gone see Harris, Discontinuous morphemes (fn. 5 above).

⁷ Pike appears to have been among the first to use the term ‘nucleus’ in this sense. See his Analysis of a Mixteco text (fn. 3 above).

⁸ The equations in Zellig S. Harris, From morpheme to utterance, Lang. 22.161–83 (1946), though not using the term IC, are based on the ‘nuclear hypothesis’. His highest surviving numbered formulas represent the basic nuclei. The eliminated formulas at each level represent the satellites.

⁹ Premises 1, 2, and 9 have been suggested, in slightly different form, by Hockett, and Premise 4 by Bloch, in Charles F. Hockett's review of Nida's Morphology, Lang. 23.273–85 (1947), esp. 282–3. Several of these premises were used by Kenneth L. Pike and Eunice V. Pike, Immediate constituents of Mazateco syllables, IJAL 13.78–91 (1947).

¹⁰ Many linguists might accept this as the only valid form of rank, exluding all exocentric constructions. But rank is nevertheless implied by the description of an exocentric form like love-ly as consisting of stem plus affix.

¹¹ A constitute is an expression that consists of two (or more) ICs; see Wells, op.cit.

¹² Martin Joos, Statistical patterns in Gothic phonology, Lang. 18.33–8 (1942), distinguishes between text frequency and list frequency. I am referring here to text frequency.

¹³ George E. Van der Beke, French word book (Publications of the American and Canadian Committees on Modern Languages, Vol. 15; New York, 1929); Bayard Q. Morgan, German frequency word book (PACCML, Vol. 9; New York, 1928); Milton A. Buchanan, A graded Spanish word book (PACCML, Vol. 3; Toronto, 1927).

¹⁴ This is probably much less likely in syntax than in morphology.

¹⁵ George K. Zipf, Relative frequency as a determinant of phonetic change, Harvard Studies in Classical Philology 40.1–95 (1929); The psycho-biology of language (Boston, 1935). Martin Joos, in his review of the latter work, Lang. 12.196–210 (1936), though rejecting Zipf's causal relationship between frequency and length, nevertheless appears to admit the correlation as a ‘functional interrelation’.