The title highlights a respect in which morphology differs from syntax. It makes sense to ask whether a certain word formation process (a particular affix, let’s say) is in or out of fashion, and self-appointed language pundits comment on such changes in linguistic fashion regularly in the media. However, nobody comments on fashions in how questions are formed, or in the structure of relative clauses, for instance. Syntax is stable in a way that morphology is not. This is surely connected with the fact that, many morphological processes are haphazardly ‘gappy’ (that is, they may not be formally general even if they are formally and semantically regular), whereas few if any syntactic constructions are ‘gappy’ in this way. In morphology, gaps get filled, or else gappy processes lose their regularity and survive only in a few lexically listed lexemes, like the process of forming abstract nouns by suffixing -th to adjectives, while other processes become increasingly regular to replace them.

A systematic study of morphological fashions belongs to a historical study of English word formation rather than to an introductory survey such as this. However, I will mention two fashions that manifested themselves in the last half of the twentieth century, because both of them, in some degree, go against more general trends of the last couple of centuries. The first is a fashion for certain Latin- and Greek-derived prefixes; the second is a fashion for a certain kind of headless compound.

Conscious borrowings from Latin and (to a lesser extent) Greek were fashionable in certain literary styles of the sixteenth and seventeenth centuries, because of a perceived need to enrich the English vocabulary. But such borrowings, often obscure and even incomprehensible to ordinary readers, were also attacked as ‘inkhorn terms’ – mere products of the pedant’s desire to show off his knowledge of Latin. The result is that the Latin- and Greek-derived element in the vocabulary of English has, since the eighteenth century, been pruned rather than increased. Histories of the English language standardly draw attention to Latin-derived words that used to be common but are no longer used, such as eximious ‘excellent’ and demit ‘dismiss’. One might have expected, therefore, that few new words formed during the last two centuries (apart from technical terms involving combining forms) would contain Latin-or Greek-derived elements. But this is incorrect. Since the nineteenth century a small countertrend has set in, involving the Latin-derived prefixes super- and sub- and Greek-derived ones such as hyper-, macro-, micro- and mega-. Words such as superman (originally a translation by George Bernard Shaw of Nietzsche’s German coining Übermensch), superstar, super-rich and supercooling illustrate the use with free Germanic roots of a prefix that was once typical with Latin-derived roots, often bound, as in supersede and superimpose. Words such as hypersensitive, hyper-market and hyperactivity (as in attention deficit hyperactivity disorder, or ADHD) illustrate a similar tendency with Greek prefix meaning ‘over-, excessive(ly)’, once peculiar to combining-form words such as hyper-trophy ‘excessive growth’. A more recent illustration of this trend ha been the extension to free roots of Greek mega-, so as to create megastore, mega-merger and megabucks alongside earlier words such as megalith and megaphone. A contributing factor, no doubt, is a desire to show one’s awareness and understanding of new technical terms incorporating mega-, giga- and nano-, meaning respectively ‘million’, ‘(American) THE HISTORICAL SOURCES OF ENGLISH WORD FORMATION 109 02 pages 001-152 18/10/01 3:43 pm Page 109 billion, or thousand million’, and ‘one (American) billionth’ (as in nanogram ‘10–9 grams’). Fashions in language are as hard to predict as fashions in clothing, but it will not be surprising if giga- and nano- soon acquire the same currency as mega-, macro- and micro-, with the meanings ‘huge’ and ‘tiny’.

Headless or exocentric compound nouns such as redhead, lazybones and pickpocket do not reflect productive patterns in modern English. It would be a rash writer or speaker who coined a word such as climbrock or long-neck, expecting the reader or hearer to interpret it unthinkingly as meaning ‘rock climber’ or ‘person with a long neck’. However, there is another kind of exocentric compound noun involving a verb and an adverb or preposition, illustrated by write-off, call-up, take-over and breakdown. Usually these can be related to phrasal verbs, such as in They wrote those debts off and He was called up for military service. However, compounds do not exist corresponding to every phrasal verb; for example, I have never encountered the hypothetical nouns ‘give-up’ ‘surrender’ or ‘put-off ’ ‘postponement’. Even this kind of exocentric compound, therefore, seems to be only marginally productive. Yet in the 1960s there arose a vogue for a class of compounds of the form V-in, such as sit-in, talk-in, love-in and think-in. What is curious about these is that corresponding to most of them there is no phrasal verb. People who had participated in a twelve-hour sit-in would be unlikely to describe what they had done by saying We sat in for twelve hours. The phrasal-verb-based pattern of headless compound thus for a while extended its scope outside the domain where it had previously been regular (although not fully general), but with its second component restricted to the preposition in. This exemplifies yet again a characteristic of morphology especially: the propensity to display random exceptions and lexical restrictions.