of the rare morphological type, they originate directly from word-combinations and are often homonymous to them: cf. a tall boy — a tallboy.
In this case the graphic criterion of distinguishing between a word and a word-group seems to be sufficiently convincing, yet in many cases it cannot wholly be relied on. The spelling of many compounds, tallboy among them, can be varied even within the same book. In the case of tallboy the semantic criterion seems more reliable, for the striking difference in the meanings of the word and the word-group certainly points to the highest degree of semantic cohesion in the word: tallboy does not even denote a person, but a piece of furniture, a chest of drawers supported by a low stand.
Moreover, the word-group a tall boy conveys two concepts (1. a young male person; 2. big in size), whereas the word tallboy expresses one concept.
Yet the semantic criterion alone cannot prove anything as phraseological units also convey a single concept and some of them are characterised by a high degree of semantic cohesion (see Ch. 12).
The phonetic criterion for compounds may be treated as that of a single stress. The criterion is convincingly applicable to many compound nouns, yet does not work with compound adjectives:
cf. 'slowcoach, blackbird, 'tallboy,
but: blие-'eyed, 'absent-'minded, 'ill-'mannered.
Still, it is true that the morphological structure of these adjectives and their hyphenated spelling leave no doubt about their status as words and not word-groups.
Morphological and syntactic criteria can also be applied to compound words in order to distinguish them from word-groups.