The word DERIVATION and its cousins DERIVE and DERIVATIONAL are technical linguistic terms. Thus they ought to be models of sober precision. But in fact these words carry several meanings and they have a way of slipping from one sense to another without warning, so that when an author uses a word from this group his meaning is not always unmistakable.

The continuous gradation of colour which exists in nature is represented in language by a series of discrete categories. Athough there is no such thing as a natural division of the spectrum, every language has colour words by which its speakers categorize and structure the colour continuum. The number of colour words and the manner in which different languages classify the colour continuum differ. Bassa, a language of Liberia, has only two terms for classifying colours; hui and ziza (Gleason, 1955: 5). Hui corresponds roughly to the cool end of the spectrum (black, violet, blue, and green) and ziza corresponds to the warm end of the spectrum (white, yellow, orange and red); in Bambara, one of the languages of the Congo area, there are three fundamental colour words: dyema, blema and fima (Zahan, 1951: 52). Dyema includes white, beige, and natural (cotton) colour; blema denotes reddish, brownish shades; and finally fima includes dark green, indigo and black. Maerz and Paul (1930) list over 3000 colour names in English, but generally it is considered to have eight basic names: black, white, red, orange, yellow, green, blue and violet.

1. The development of transformational theory has been marked by a considerable shifting about of powers allotted to the various constructs. As a theory of lexicon arose, for example, some of the function of the context-sensitivity of the phrase-structure rules was rendered unnecessary; correspondingly, when the rewrite system was tightly confined to a context-free nature, the transformational component took up part of the burden of providing for certain dependencies and concord phenomena. A fair estimate of the history of that component would have to concede that its powers – in spite of the cycle, and some scattered efforts (Emonds, 1969; Postal, 1971; Ross, 1969; Sanders, 1970)–have grown significantly. For example, Emonds's work aside, no general principles of derived constituent structure have developed (as originally anticipated, say, in Lees, 1957b: 400–401); quite the contrary, the particular elementary operations at the root of transformational relationships have been extended, so that at the moment, in addition to sister-adjunction and daughter-adjunction, we have at our disposal Chomsky-adjunction – a range of moves allowing just about any sort of bracketing relation to be developed. Similarly, the recent suggestion that a transformation be considered, most generally, a relation holding between P-markers (not necessarily ‘adjacent’ and not necessarily paired) engenders an enormous increase in power – all the more so since the so-called local and global constraints that are intended to offset the magnification of power are only promissory (Lakoff, 1970).

1. I want in this paper to draw attention again to the theoretical basis of the use of distinctive features in generative phonology, and with that in mind, to consider what improvements can be made. One function of distinctive features is to provide a formal means of expressing the notion of a NATURAL CLASS (cf. Harms, 1968: 26), such that a phonological rule which applies to a natural class of segments may be expressed in a simpler way than a rule applying to some other class of segments. Distinctive features, and notational devices and conventions, are intended to capture this notion of simplicity so that by examining a formalized rule one may discover, counting the symbols in it according to a set of values provided, at least whether the rule is simpler than a comparable rule, if not its simplicity in any absolute sense. It is also intended that the set of features should be a substantive universal of language, though, as Harms' examples (1968: 23–38), drawn from various people's analyses of various languages, show, it cannot yet be seen that there is a tendency towards agreement on which these might be, as Harms himself (38) is aware. He sees the problem, however, in these terms:

The basic set of features can be viewed as a hypothesis about language, subject to empirical validation. Arguments for adding new features to the list or for altering the basic features must demonstrate the inadequacy of the basic hypothesis. Such arguments, however, cannot be based upon an appeal to the simplicity metric, for the features themselves are elements in the simplicity metric. Phonemic solutions that do not assume the same set of features cannot be compared in terms of simplicity. The claim must be that one solution provides a more reasonable hypothesis of the phonological structure of the language in question.

In several recent articles the issue of directionality in transformational grammar has been treated, rather unsatisfactorily to my mind. The question is this: are the relationships among the various levels of grammatical description (semantic structure, deep structure, surface structure, phonetic structure) such that certain levels are descriptively prior to others? That is, is there an inherent ‘direction’ to the relationship between two levels of description (say deep structure and surface structure)? Recent treatments suggest that the question is pointless, or that the answer is no. I maintain that this impression results entirely from the way previous discussions have been worded, and that the issues have yet to be approached properly.