Oxford University Press has recently published scholarly editions of two books by William Empson (1906-1984): Some Versions of Pastoral first published in 1935 and now running to 496 pages and £80 in Seamus Perry’s edition; The Structure of Complex Words, originally 1951 and now running to 672 pages at £95 in Helen Thaventhiran and Stefan Collini’s edition. Colin Burrow, Professor of English and Comparative Literature at Oxford, provides a long and informative review of the books in my London Review of Books 15 July 2021. But he hasn’t quite persuaded me to place an order.
Empson is reckoned a major figure in the development of literary studies in England and the USA, a position achieved despite being expelled from the University of Cambridge for heterosexual activity (see Wikipedia for details). It is his role in the development of modern literary studies which justifies these fresh editions, not his status as persecuted heterosexual. Leaving aside the first book it is the title of the second, The Structure of Complex Words, which deters me.
That invites two questions: (1) What kind/s of structure do words have, if any? (2) Unless the expression Complex Words is a pleonasm, what distinguishes complex words from the rest?
Words when spoken have phonetic and phonological features which can be captured by the phonetic alphabet; when spelt out in an ordinary alphabet some of those features may be recoverable from the written form but in English not consistently so. So the pronunciation differences between single-syllable hat and hate are indicated by application of a rule which extends to other cases: fat and fate, mat and mate, pat and pate, rat and rate … Unfortunately, there is also bait as well as bate, gait as well as gate. And at the outer limits of English orthography, sound and spelling are entirely separated, notably in the case of proper names of which the stand-out (and meant to be stand-out) is Featherstonehaugh: the idea here is that only if you mix in the right circles will you know how to pronounce it. So whatever structure there is, it is not built out of Lego blocks which retain their shape wherever inserted.
If hat and hate are simple, are hatful and hateful complex? Well, they are two-syllabled rather than one; the bolted-on ending - ful adapting them to be ready for adjectival use. That doesn’t sound very complex, and it’s not hard to grasp. But sometimes the - ful is not - or is no longer - a purely grammatical indicator. I can look at the night sky and tell you that I experienced a moment of awe but if I tell you that I had an awful day it’s not in the same league of meaning; the - ful downgrades awe to something more quotidian. Someone learning English as a foreign language might have to have that explained to them. Indeed, whatever complexity words have out of context of use might best be understood in terms of the problems they might pose a second language learner. The Oxford English Dictionary is constructed very much in this spirit though it also has English speakers in mind when it classifies words as arch., euphem., obs., coarse slang. In this way of looking at things, it’s not clear to me that the complexities of awe and awful are of a different order to those which surround, say, boat and ship. I don’t believe that the latter pair could be handled as a “technical” problem, solved by recourse to some measuring system, in a way that awe and awful can’t.
What could be structure turns out to be a dense mass of sense congealed from a long past of history and cultural changes. History and culture do not have modular structures, as if atoms or logical formulae or those Lego bricks. Accurately summarising such history and culture within the pages of a dictionary is a Sisyphean task because history is accumulating as you write and culture is changing in new ways ditto. Dictionaries are just specialised encyclopaedias, and thus necessarily partial. The rules of arithmetic or the periodic table aim to escape partiality and many believe that they can succeed.
But are there other kinds of complexity? Is hate a complex word in ways which hat isn’t? Is onyx simple in a way which honest isn’t?
Speakers and writers are, for the very large ninety nine point nine percent most part, using words which have been used before and where there are some fairly general features they will be understood to possess or imply unless those features are explicitly cancelled. So if I say She was wearing a hat that implies (whether I have realised or not) something which may occasionally be cancelled as it is in She was wearing a hat slung round her neck and hanging down her back. Ah, you see it now, a straw sun hat, perhaps? If I say He’s a hateful fellow that may be partially cancelled by addition of the very convenient but word He’s a hateful fellow but I have to give him credit for ….
This last example incidentally illustrates another problem for any theory of structure: hateful was once a twin of hatful, since it meant full of hate. But that use is now reckoned archaic by the OED and hateful is now roughly synonymous with repugnant - a hateful person is one who excites hate not someone who exudes it. There is no logic or necessity requiring this change, any more than there is for the replacement of reject with refute or annoy with aggravate. This kind of complexity in the history of individual words makes the prospect of any general account of their "structure" very remote. It can only be captured by narratives. The possibility of a more general theory of how to use words in a particular context is much more promising as is shown by the world-wide take-up of J L Austin’s How to Do Things with Words (1962) and all that has been developed from its simple expositions.
A writer needs to be sensitive to the sense which words are likely to carry (as it were, directly) and imply (as it were, indirectly). But it’s also true that since writers cannot control everything they are doing, some senses will be conveyed and some implications steered towards which they did not, personally, at the moment of writing, intend. Forensic analysis of literary texts, under many different names, will be on the look-out for such senses and implications, perhaps especially if they seem to contradict avowed intentions which the author has been foolish enough to profess. It’s better to acknowledge that a lot of the time you ( and me) don’t know exactly what we are doing and at the very moment when we are trying to bend language to our purposes the meanings of what we say and what we write exceeds whatever we intended.
But thinking like this, what place is there for a special category of complex words? Hat has a history just as much as hate which is to say that both have been framed in different ways in different times and places. It may be fashionable or unfashionable to wear hats; ditto for hating to which in some times and places cultural prestige may attach so that it is a compliment to call someone a good hater.
There is a further problem. Words can be used but they can also be mentioned. Traditionally, there were words which it was forbidden to use but which it was permitted to mention - that is to say, to quote. But in my Brave New England Puritan print culture, there are words which it is regarded as unacceptable either to use or quote. Thus, when my Prime Minister (in a text message) described his Health Secretary as “totally fucking hopeless” both print and online versions which supposedly quoted what he said actually amended it to include asterisks which were not present in the original. The purpose is the same as that intended by Sunday school rules for turning Direct Speech into Indirect Speech, rules which miraculously transform the wine of “He’s totally fucking hopeless” into the water of “The Prime Minister expressed exasperation with his Health Secretary”. The trouble with direct speech is that it can be too direct by half and not something you would want to put on the front pages or allow your maiden aunt to hear. (Maiden aunt is probably now arch. but you can always look it up).
Part of the writer’s opportunity consists in the ability to make use of this difference between use and mention but sometimes leaving it unclear whether something is being said or quoted; access to this linguistic resource is central to making the most of irony which can be administered in larger or smaller doses, according to taste and malevolence.
Well, if I am led to generate 1500 words just contemplating the title of a book, I think it would be a very bad idea to try to read its 672 pages and distill them for you here - and Colin Burrow in his helpful review indicates that distillation isn't exactly easy; Empson wasn't that kind of thinker. But perhaps I have done enough to prompt you to fork out for this many-paged, expensive edition of William Empson’s The Structure of Complex Words.