^{page 327 note 1} I should like here to express my very deep gratitude to Mr. Liou for his help and patience during these last twelve months. I should also like to acknowledge my indebtedness to Dr. S. A. Birnbaum for having suggested the term ‘investigation’.

^{page 327 note 2} Thus an isolate like shanq may occur in any one of four categories described in this paper because it may keep the same ‘grammatical company’ as the isolates in any one of the four categories. In a form like [shanq=i-tsyh], for instance, it will be described as a ‘numeral prefix’ and grouped with other isolates such as dih in [dih=i-tsyh] ‘the first time’ and tour in [tour=i geh reri] ‘the first person’. In a form like [shanq-yueh] ‘last month’ it will be described as a ‘specific determinative’ occurring in an ‘unmarked specific form’ and grouped with isolates like jin in [jin-nian] ‘this year’ and shiahshiah in [shiahshiah-shingchyi-ell] ‘Tuesday fortnight’. In a form like [shanq-tour] ‘top’ it will be described as a ‘specific determinative’ occurring in a ‘marked specific form’ and grouped with isolates like dong in [dong-buh] ‘Eastern part’ and yow in [yaw-miann] ‘right’. In a form like [juotz-shanq] ‘on the table’ it will be described as a ‘substantival suffix’ and grouped with isolates like lii in [utz-lii] ‘in the room’ and yiiney in [san-nian-yiiney] ‘within three years’. Analysis of a piece like shanq chwan ‘to embark’ falls outside the limits of this paper, but shanq would there be described as a ‘verb’ and grouped with isolates like shiee in shiee tzyh ‘to write’, whilst in a piece like nii shanq naal chiuh ‘where are you going?’ it might be described as a‘co-verb’ and grouped with isolates like daw in woo daw Beeijing chiuh ‘I'm going to Peking’.

^{page 327 note 3} cf. Togeby, K., Seventh Intern. Cong. of Linguists, Prelim. Reports, p. 59Google Scholar, ‘De même qu'on peut classer les particules d'une langue flexionelle d'apres leurs possibilités de combinaison avec les autres éléments de la langue, on devra pouvoir caractériser les mots invariables d'une langue isolante’ (my italics).

^{page 328 note 1} Firth, J. R., ‘Modes of Meaning’, Essays and Studies, 1951Google Scholar, The English Association, pp. 123 ff.: ‘The statement of meaning by collocation and various collocabilities does not involve the definition of word-meaning by means of further sentences in shifted terms. Meaning by collocation is an abstraction at the syntagmatic level and is not directly concerned with the conceptual or idea approach to the meaning of words. One of the meanings of night is its collocability with dark, and of dark, of course, collocation with night’.

^{page 328 note 2} Throughout this paper square brackets will mark a complex form.

^{page 328 note 3} Chao, Yuen Ren, Mandarin Primer, Harvard University Press, Cambridge, Mass., 1948—CrossRefGoogle Scholar hereafter cited as Mandarin Primer—pp. 39–59.Google Scholar

^{page 329 note 1} Mandarin Primer, pp. 45–7.Google Scholar

^{page 329 note 2} Thus it is part of the grammatical meaning of verbs like gau ‘to be tall’ that they may not be directly followed by a form of the noun complex but may be followed by ‘non-specific forms’ of the determinate complex, e.g. gau i-deal ‘a little taller’, gaule san-tsuenn ‘three inches taller’.

^{page 330 note 1} cf. p. 352 below.

^{page 331 note 1} Yoshitake, S., ‘The Grading Method of forming Numerals’, T.P.S., 1940, p. 53.Google Scholar

^{page 331 note 2} It is to be noted that units have no colligability with nouns.

^{page 331 note 3} Shyr as the initial part of a compound numeral determinative, e.g. shyr-san ‘thirteen’, does not need to be preceded by a unit, but in non-initial position it must be preceded by a unit, e.g. ibae-ishyr-ell ‘a hundred and twelve’. It may also occur as a single numeral determinative which has colligability with determinators, determinates, and multiples, and is therefore listed both as a unit and as a multiple.

^{page 332 note 1} There appears to be no difference in meaning or usage between the forms jeh and jey, nah and ney, and naa and neei. Only jeh and nah, however, occur as ‘non-complex units’, e.g. jeh sh sherme? ‘what is this?’

^{page 332 note 2} Of the two isolates meei is used with greater frequency than geh. Chao distinguishes between them by translating geh as ‘the various’ and meei as ‘each’, cf. Mandarin Primer, p. 51. 29.Google Scholar

^{page 332 note 3} cf. p. 346 below for a discussion of the form Dave Noun.

^{page 333 note 1} In non-mathematical language ell occurs in compound numeral determinatives only, where it is used instead of leang in end-position and also before the multiples shyr and bae, e.g. ellbae-ellshyr-ell ‘two hundred and twenty-two’. Leang is normally used before chian and wann, but ell also occurs. Ell is never used directly preceding a determinator or determinate, except in literary style.

^{page 333 note 2} cf. note 3, p. 331, above.

^{page 333 note 3} cf. also p. 331 above.

^{page 333 note 4} In the material examined the quantifier ju occurred only in colligation with the generic determinator wey, e.g. ju wey torngshyue! ‘fellow students!’

^{page 334 note 1} The determinate complex [ibae-nian-chyan] is here a form of the larger noun complex [ibae-nian-chyan-de shyhchyng] [Deseg Noun].

^{page 335 note 1} These two examples are an attempt to state the ‘meaning by collocation’ of the two pairs chyan= how= and shanq= shiah=.

^{page 335 note 2} cf. p. 352 below.

^{page 335 note 3} diishiah ‘underneath’ is used more frequently than the specific form shiah-tour.

^{page 336 note 1} cf. p. 352 below.

^{page 337 note 1} Yeou san-chyy charng and yeou san-jin jonq are major colligations where neither chyy nor jin have colligability with a noun and where they must, therefore, be described as determinates, cf. also discussion of the form Dave Dor, two paragraphs below.

^{page 337 note 2} cf. also ‘restricted determinates’, below.

^{page 338 note 1} cf. also ‘substantival suffixes’, p. 343 below.Google Scholar

^{page 338 note 2} Chao distinguishes i juotz tsay ‘a tableful of dishes’ and i juo tsay ‘a regular set of dishes forming a dinner for one table’, Mandarin Primer, p. 46.Google Scholar

^{page 338 note 3} cf. below p. 348.

^{page 338 note 4} Simon, W., A Beginners' Chinese-English Dictionary, Lund Humphries, London, 1947, pp. xlii ff.Google Scholar, ‘The Most Important Classifiers’.

^{page 338 note 5} Chao, Yuen Ken and Yang, Lien Sheng, Concise Dictionary of Spoken Chinese, Harvard University Press, Cambridge, Mass., 1947, pp. 283 ff., ‘List of Auxiliary Nouns (AN) and Units of Measurement’.Google Scholar

^{page 340 note 1} They are thus distinct from ‘verbal de-segments’ like pao-de in to pao-de tay kuay ‘he's running too fast’. In a verbal de-segment the particle is preceded by a verb and followed by a form of the verb complex.

^{page 343 note 1} The material examined showed no instance of a suffix following a marked specific form of the determinate complex.

^{page 343 note 2} The suffix shiah is not common. Mr. Liou almost invariably uses the disyllabic suffix diishiah, e.g. juotz-diishiah ‘under the table’. Jong in kongjong ‘in mid-air’ cannot be regarded as a substantival suffix since there is no substantival form kong.

^{page 343 note 3} This form appears to be much more common than the form Deseg SDave-Date (jey geh ren-de howtour). cf. p. 352 above.Google Scholar

^{page 344 note 1} A post-verb being an auxiliary verb following the main verb in a form of the verb complex like getzay. The ‘suffix form’ may also occur after tzay or daw when these are the only verbs in the colligation, e.g. shu tzay juotz-shanq ‘the books are on the table’.

^{page 345 note 1} Both Chao—Mandarin Primer, p. 34—and John De Francis—Beginning Chinese, Yale University Press, 1946, p. 46—have defined nouns in a similar way.

^{page 345 note 2} cf. p. 338 for a discussion of the form Dave Dor (jeh been) which does not include a noun but is nevertheless regarded as a form of the noun complex.

^{page 346 note 1} cf. also p. 347 below.

^{page 346 note 2} cf. p. 331 above.

^{page 347 note 1} cf. also Mandarin Primer, p. 51Google Scholar, ‘after a verb i is often omitted’.

^{page 347 note 2} See examples, p. 349 below.

^{page 347 note 3} Mandarin Primer, p. 51. 28Google Scholar

^{page 348 note 1} Mandarin Primer, p. 39. 12.Google Scholar

^{page 351 note 1} p. 337 above.

^{page 351 note 2} Any one restricted determinate normally has colligability with only one numeral determinative.

^{page 352 note 1} cf. also note 3, p. 343, above.

^{page 352 note 2} cf. Mandarin Primer, p. 205Google Scholar, Gangx tzay [ney-kuay yuntsae-de shangtour-yowbial] ‘just above and to the right of that cloud’.

^{page 353 note 1} cf. p. 347 above.