^{page 534 note 1}In his book Studies in Vietnamese (Annamese) grammar, University of California Press, 1951, p. 44, 2.1Google Scholar, Professor M. B. Emeneau emended the definition of the Vietnamese word slightly when he stated :

‘The basic unit of the syntactic analysis of the language is the word, which is the phonological unit and, at the same time, the morphological unit as was stated in 1.1. The word and the morpheme coincide, with the qualification made in 2.4.’ The qualification referred to was concerned with reduplication in groups of two morphemes where there was a partial similarity of form between the two.

The Vietnamese scholar Dr. Lê văn Lý (Lê purler vietnamien, Éditions Huong Anh, Paris, 133), redefined the Vietnamese word and established two types of word, ‘mot simple’ and ‘mot composé’. The mot simple is defined as ‘Un signe vocal dont la formation peut partir d'un simple phonème ou de la combinaison de plusieurs phonemes, dont l'expression vocale se fait en une seule émission de voix ou une seule syllabe que l'écriture présente par une unité distincte et qui offre un sens intelligible’.

The mot composé is defined as a unit of more than one syllable, substitutable for a mot simple, and expressing a single concept.

It is clear from this that neither Emeneau nor Lê văn Lý was satisfied with the usual definition of the Vietnamese word.

^{page 535 note 1}Emeneau did attempt such a classification (op. cit., p. 79, 2.14) because, according to his definition, such units are words, but the following quotation shows the difficulty into which such an attempt leads.

‘An example is hâu, which occurs only as an attribute following the noun du-a which denotes any species of cueurbitaceous plant, including various types of melon and cucumber. The phrase “dua-hâu” means “watermelon”. It is impossible to assign hâu to the verb or to the noun class, and the decision can be suspended. The parallels dua-chuột “cucumber” in which chuột is the noun “rat”, and du-a-do, also “watermelon”, in which do is the verb “to be red”, hardly help the analysis.’

^{page 535 note 2}Some of the criteria used in the classification of words described in this paper resemble those set out by Dr. Lê văn Lý in his chapter dealing with word classification (op. cit., 145). The Vietnamese word as defined in this paper is very different from the word defined by Dr. Lê văn Lý. The number of word classes established in this paper and the content of these classes are different from those of Dr. Lê văn Lý.

^{page 537 note 1}I have adopted the term ‘indicator’ from the language of chemistry in which it describes a reagent used to indicate by change of colour, the presence of an acid, alkali, etc.

^{page 539 note 1}For ‘colligation’ see Simon's, H.F. article ‘Two substantival complexes in Standard Chinese’, BSOAS, xv, 2, 1953, 327–55.CrossRefGoogle Scholar

^{page 543 note 1}It is necessary to point out that difficulties may appear to arise in the case of homophones. For example, in the following utterances:

Dê cái cu-a này trên eái bàn.

place, object, saw, this, upon, object, table.

Put this saw on the table.

Tói phai cura nhiều cêy trong khu này

I, must, saw, many, tree, within, region, this.

I have to saw down a large number of trees in this region.

the unit ctra must be assigned to two different word classes. In the first utterance cvra directly precedes the indicator nay and is assigned to word class 3. In the second utterance cira directly precedes the indicator nhiêu and must be assigned to word class 2. It is stated in all such cases that the unit in question represents more than one Vietnamese word, and that these words are homophones. In the examples given, the unit cua represents two Vietnamese words. Indeed, the English translation of cua, namely ‘saw’, may be a member of both nominal and verbal categories: to saw: the saw.

A further parallel from English may serve to exemplify this apparent difficulty. It is possible to distinguish homophones in English by means of scatter. The English noun ‘ride’ has a scatter of ‘ride, rides’, and may be distinguished from the English verb ‘ride’ which has a scatter of ‘ride, rides, rode, riding, ridden’. Vietnamese is not an inflecting language and it is not possible to distinguish Vietnamese homophones by this criterion of scatter.

^{page 543 note 2}Dr. Lê văn Lý pointed out the advantages of such a method of word classification (op. cit., p. 146, 1.2), but himself resorted to subjective criteria in the establishment of his categories as the following extract shows:

‘Sont identifies comme A tous les mots précédés d'un classificateur tel que cái, con, nguời, ou d'un mot générique comme ké, sự, dồ, việc, nghẽ.’

The categories classificateur and mot générique have nowhere been established by Dr. Lê văn Lý and it is very doubtful whether any two scholars would agree on the question of which Vietnamese words are classificateurs and which mots génériques.