Wednesday, April 11, 2018

In defence of would of

A pet peeve of mine is those people who enjoy snobbishly correcting other people's English on the interwebnet. Which might surprise you, since I have made an astonishingly poor living out of correcting English. But then, I have also taken a shit once every day and that doesn't mean I'm keen on doing it in the high street.

I want to talk to you about could of, would of, should of, which people spend far too much time getting all snobby about. Not only do those people make an ugly spectacle of themselves, but they are wrong. Yes, really.

Three things back my play here: how we learn English, the variety and mutability of Englishes and why pedants are wrong about how language obeys "logic".

Babies learn how to speak and they're very good at it. Something that we often forget is that brains are much smarter than people. Weird thing to say but trust me, it does make sense. Whatever difficulties we seem to have with processing things in our world, our brains deal with extremely complex and large amounts of data. Perhaps not always perfectly but quickly and quite efficiently.

So babies are able to sift through sounds in the world around them to pick out, analyse and categorise ones that form language. They are able to internalise huge bodies of rules, which are in most languages complex and difficult enough that learning them for adults can be trying, and they're able to exclude sounds that don't fit on account of not being part of language at all or of belonging to another language.

But what's important to remember here is that they store language as a system of sounds. Babies do not learn to write. Our formal education in literacy largely consists of marrying sound to written symbol. And it's interesting that babies can learn to speak so well but children can learn to write so badly.

I can hear our pedant object. But people don't speak so well! How can you say such a thing? Some of those people speak common. Well, here's the first place they are wrong. You need to ask yourself what a baby is trying to do. It's not trying to learn to speak English. It has no idea what English is. It does not aim for a platonic ideal of a language. It is trying to learn to be comprehensible to the beings it finds itself among. It is not even trying to understand them. Babies are pretty much autistic; they have no interest in what you want or what you're trying to say. They are interested in getting what they want, and language is a good way to help them get it.

Babies are very good at learning an idiolect that can (eventually) be understood by other nearby speakers. (An idiolect is your own peculiar way of speaking your language. We all have a different idiolect. It's our most natural way of speaking. You can have more than one, of course, because you have a separate idiolect for each dialect you speak -- but this is verging on a technical morass because these are only terms of art for what we have: people speaking utterances that they hope other people will understand.)

Part of how they do it so well is they form rules that they can generalise across different utterances (I'm going to continue to use this word from linguistics to distinguish spoken sentences from written ones). And anyone who has a child knows they overgeneralise. Which can be very cute. Who doesn't love it when a child says Mummy drived the car or Daddy goed to work? All the child has done is apply the rule "form the simple past by adding /t d/ to the verb stem" too broadly. (Yes, I know the rule is a bit more complex but that's not the point.)

Children's brains are searching for patterns in the sounds they hear. They are looking for rules that apply systematically. This is because language is "generative". This quality means you can create novel utterances from the "dictionary" you've acquired. And you can do it from whatever materials you have. Children are very creative with language. As soon as they have acquired even a few words they try to express themselves with those words. The rules help them do that by providing structures to hang the words on.

Now note that the rule we have mentioned involved adding a phoneme to the end of a word. (A phoneme is a distinctive sound in English -- in fact, the "rule" is something like "add the phoneme /d/ to the end of the lexeme while observing phonotactic rules". And the marker for the past tense can be /d/ or /t/ without changing meaning -- there are phonological rules for which one to use.) Park that while we take a trip into history.

Old English -- Anglisc -- was a synthetic, inflecting language. Synthetic is a pole on a spectrum; the other is analytic, and the spectrum is a description of how much meaning there is in each "word". So an analytic language such as Chinese has monosyllabic words that have one meaning each (you create more meaning by positioning the words) and a (poly)synthetic language such as Navajo has multisyllabic words that contain a lot of meaning (they can be as long and meaningful as an English sentence yet are structurally one word). Inflecting means that changes in meaning or relationship between words in a sentence can be shown by changing parts of the word, usually the ending, without adding more elements (contrasting with agglutinative languages, which tend to add new syllables for more complex words, in which syllables are fixed -- Turkish can use five or six syllables where Latin would use one or two).

Anglisc used cases to show relations between words in a sentence and inflected verbs to show differences in tense, aspect and number. Anyone who's learned Latin will be familiar with the concepts of case and inflection. Over time, however, Anglisc evolved to become more analytic. Relations between words became more often shown by prepositions or word order, and verbs became much less inflected. But the new analytic English was littered with fossils from its synthetic past. For example, we have a fossil of Anglisc genitive case in possessives. Where Anglisc had "dog" and "dogges", we have "dog" and "dog's".

In fact, the possessive is the only case remaining for English nouns. Importantly, it cannot be seen as inflectional. The rule is simply to add the phoneme /s z/ to the end of words (with some jiggery pokery and exceptions of course).

Now imagine you are a baby presented with a jumble of sounds. You are able to categorise, say, "dog" and "dog's" and you systematise the difference between them and other words used in similar contexts as a rule to add a phoneme. In other contexts, other words seem to add other phonemes. This seems to be what English does.

Interestingly, it really is what spoken English does. The baby is presented with (forgive my "phonetic" spelling): /ai/ /yu/ /hi/ /wi/ /yu/ /the/ and /aiv/ /yuv/ hiz/ wiv/ /yuv/ /thev/. The baby's already systematised that /hi/ /shi/ /it/ are exceptions, so there's no problem with simply assuming "add v" is the rule for these words to express something. But what?

Well, "I've done" is the perfect. This is how we express an action that is completed (perfected) from the point of view of the present time. You use the auxiliary "to have" and the past participle.

In most varieties of spoken English, the auxiliary "to have" is expressed by the phonemes /v/ and /z/. In fact, /v/ and /hav/ are not interchangeable in most varieties -- they have strictly delineated contexts. They are to most intents and purposes different words. "I've done it" and "I have done it" do not actually mean the same thing for most of us. In some ideal version of English, "I've" is "short" for "I have" but babies do not learn that. They learn /aiv dun/ is the perfect of the verb /du/.

Importantly, in most dialects, the phoneme /v/ cannot be used for /hav/ in every context. Some dialects do allow /aiv thri/ for "I have three" but curiously they do not allow /hiis thri/ to mean "he has three". Generally, /v/ is not used for /hav/ though, and in spoken English, it's easy to schematise it as a postfix that allows formation of some tenses and/or aspects.

Spoken English does not have spelling. Not only do children not store the words they learn as spelled versions, the spoken language cannot be understood as identical as the written language. I'm not going to go into great detail about how this can be known: I'll simply appeal to your own intuition and the fact you cannot spell every word you can say (don't lie, you can't).

Given this, it's not clear that there's a "proper" way to spell /wudv/. By convention we make out it's a contraction of "would have" but children do not learn "would" and "have" and "would have" and "would've" as connected in any vital way. I'm just asserting this btw; I don't propose to provide evidence for it.

We have talked about how children develop an idiolect that sufficiently corresponds with what people around them speak that they can be understood. Then they go to school and if like me they come from an area with a slightly different dialect, or indeed they are from somewhere diglossic, they find out how they speak is "wrong". I think for English speakers it's fairly hard to understand that written standards are somewhat idealised versions of the languages they represent. If we look at Arabic, it becomes much clearer. Written standard Arabic does not bear much resemblance to vernaculars in Morocco or Iraq, and speaking it is a bit like declaiming Chaucer in Chaucerian English -- you need a special and distinct education to understand it. Chinese is worse! I've mentioned in other posts that Cantonese, for instance, is written in the same script as Mandarin, but they are very different languages -- perhaps as far apart as English and Italian as far as vocabulary goes.

And one thing we learn is that /wudv/ is represented in the written standard language as "would've". Which makes it look like something's missing. But of course it isn't.  I think that what is going on is clearer when you compare English with a Brythonic language, where there are different ways of rendering the present tense (roughly a personal and an impersonal way). There are informal and formal ways to render the past modals: /wudv/ and /wud hav/, /kudv/ and /kud hav/ and so on (I'm not ignoring the variety /wud@/ -- it's just not useful in this discussion, although it will rear its ugly head later). I will note that some people eschew contractions, preferring to represent everything, even in dialogue, in good written language. I respect that choice because it avoids making uncomfortable choices with dialectal variation. Nothing says that written English need be phonetic or that one cannot understand that a character uses a contraction even though it's not written.

All of this is a long path to saying that "would've" has shaky legitimacy. It renders a spoken construction that arguably does not represent a contraction at all. And now to our point. The phoneme /v/ serves in the spoken language as the preposition written "of". It's so commonly used that one should probably consider /ov/ to be a variation on it, not the other way round. It would be beyond pompous to say /bag ov tschipz/ for instance. And this word is uncontroversially represented by the written word "of".

I said we would mention again /wud@/ but actually let's look at the purposive /aim gon@/. Again, outside emphatic uses, it's very common for "going to" to be pronounced /gon@/. Writers have long affected to write this as "gonna". And to be honest, if someone wrote this in informal English, it'd be a brave pedant who objected. Furthermore, the cliche "shoulda woulda coulda" would be rendered ridiculous were it spelled "should've, would've, could've". Why? Because it doesn't look right.

A speaker of a language does not speak one language -- even if they are monolingual. They speak a range of lects that vary according to the person they are talking to and the context they are talking in. You don't speak "posh" to your mum. Most speakers of most languages actually can and do use an entirely different dialect at home than they do in the public sphere, although in English, dialects tend to be so close to standard English that you cannot fairly say that. Still, even though we may use the same variety of English in different settings, we use different vocabularies and different ways of expressing ourselves. These different sets of English are called register. I don't think anyone could argue that the "interview register" is different from the "spouse register", at least for most of us!

As an aside, the differences between register can be quite obvious but also quite subtle. For instance, if I was being interviewed, it's very unlikely I'd use vulgar language. I wouldn't say my last job was "shit" or my last boss was an "arsehole". But I would also likely say /wudv/ not /wud@/, which is what I'd normally use talking to a friend. These are examples that illustrate something we all understand: you use language appropriate to the setting. But what's not so obvious is that written English also has registers.

I don't know about you but I don't worry too much about how literate I am in a text message. And I might even use a lol here and there. (I hope that if S reads that she'll crack a smile.) Now I wouldn't say lol in a newspaper article. And I'd probably avoid difficult words and certain constructions in the newspaper too. I generally write with a fairly restricted vocabulary for various reasons but if I was writing something highly formal, I might deploy more impressive words (I might not though; I don't command a very impressive vocabulary at all).

I think even the worst pedant among us can happily agree that there are different written registers and that whereas you might write an email to a friend in a chatty tone, you'd use something more formal for business, for instance. You might not. You might write your business proposal in a lighter way if that was appropriate to the business. But you'd be aware that you were doing that.

So people writing posts on Facebook are not bound to stick to the register that would be appropriate to a book about astrophysics. Facebook would be even duller than it already is if they were. And those of you who enjoy writing that "Trump is a dick" would not be permitted to, since calling people dicks is on the whole inappropriate in textbooks. I think that a variety of what you might call "common", "vulgar" (in a more technical sense) or "vernacular" English would be most people's idea of what's appropriate. In this version, some things that I would not allow in formal English are common: objective "who" (should be "whom" but even I don't usually write that), unpossessed gerunds (Fowler would spin in his grave at you missing that), even existential or presentative singular "there's" for plural referents ("there's three good ideas about that" or "here's three things you need to know") and so on.

So here we have it.

I think "would of" is an understandable way to render /wudv/, particularly for people who are used to writing "SMS English", where apostrophes are often eschewed. (Wouldve looks weird, you know it.)

I think furthermore "would of" is a more natural way to render /wudv/ because there is no other context in which the phoneme /v/ is rendered "have" outside tense/aspect forms, and "'ve" cannot be used for "have" more generally, whereas /v/ is rendered "of" routinely.

I think we all agree that contextually "gonna" is the preferred rendering for /gon@/ and writing "going to" would not just be wrong but would be ridiculous.

I think that the settings in which people wish to disallow "would of" are not formal, so rules of formal language cannot be applied. In fact, the pedant is left in a bind, because you do not write out contractions in other cases, yet "would've", "could've" and, omg I'll shit, "hadn't've" are horrid.

I do not have enough lunchtime to discuss the truth that language does not sit still, that it changes, and its form depends on its use. Chaucer used the phoneme /n/ fairly liberally as a prefix to negate verbs. But you can't do that now. Eventually, everyone will write "would of" and it will be as though that was always how it was written.

English is not fully analytic. We don't add the meanings of "would" and "have" together. "Would have" has a synthetic meaning that is not like either of them. "Have" is definitely synthetic -- used across the Indo-European group without actually meaning anything in itself. And there's no good reason, really, for spelling it in the same way as the verb of possession.

So, pedants, you will be wrong for sure in maybe a hundred years. And mostly, you are wrong now. You are a sad lighthouse of formality in an informal world. I pity you, fighting as you do to force a tool of immense flexibility and expressiveness into dull regularity. Language is how we get others to understand us. We learn it because we want others to understand what we want, what we need, what we feel and in time to express that we understand what others want, need and feel. And you understand perfectly well what "I would of told you if I knew" means. So you are wrong, sorry.