Lexicography FAQs: spelling variants

English spelling isn’t officially standardized. Unlike French, with its Académie Française to make pronouncements on correct (and incorrect) spellings, English spelling is largely governed by consensus. Whatever spelling of a word is most widely used by a particular speech community becomes the norm.

Spelling variation in English

Of course, the vast majority of common words tend to have a single agreed spelling; we all spell house and rabbit and oxygen and establishment in the same way. There is a wonderful degree of variation though too. Some words have different spellings in different places, the most obvious examples being British and American regional spellings of words like colour/color, catalogue/catalog, tyre/tire. Other variants have regional leanings but are actually less clearly divided; encyclopedia/encyclopaedia, foetus/fetus. Then there’s the question of whether to hyphenate or not (a topic I’ve written about before). Sometimes that’s about the evolution of a word; E-mail became e-mail and has now mostly become email. Sometimes there are grammatical reasons for a difference – a part-time job versus working part time – if you’re especially fussed about that kind of thing. And sometimes it’s down to no more than personal preference. I’m more of a co-ordinate kinda person just because I think the double O in coordinate looks funny. FYI, a quick corpus search shows that I'm very much in the minority on that.

Dictionaries and spelling variants

But surely you can check the “correct” spelling of a word in the dictionary, I hear you cry. Well, first off, and for the umpteenth time, there’s no such thing as the dictionary. There are many different dictionaries which will all have slightly different policies about spelling variants, and which will occasionally – shock, horror – show different things. And that’s because as lexicographers, we don’t decide what the correct spelling of a word should be, instead we use corpus data to help us reflect how a word is spelled out there in the wild. How much of that variation we reflect will depend on the policy of the dictionary. So, for example, a learner’s dictionary is more likely to keep things simple, only showing the most frequent spelling, with maybe one variant spelling if it’s very common. Larger reference dictionaries are more likely to show a wider range of variants.

And of course, different dictionary publishers use different corpora, so especially where it’s a close call, one corpus might rank one spelling as more frequent and so a lexicographer might make it the headword (the word in bold at the top of the entry) and make another the variant (usually shown below, perhaps in brackets and labelled as ALSO). Another corpus might come up with the opposite stats and result in a subtly different entry. 

The curious case of dogtooth

One challenge for lexicographers is how far to go with variants. Last week, I started researching the word dogtooth. I was initially looking for the anatomical sense to refer to a human tooth, also known as a canine. As I looked at the corpus though, I came across lots of examples referring to a fabric pattern:

 

And to a type of violet also called a Erythronium – which is rather beautiful, and I happen to have in my garden (see below).

So far, so good, until I realized that at least for the second two contexts, I’d probably want to add a possessive S in the middle. So, I went back and widened my search criteria. I won’t share the stats from the publishers’ corpus (for reasons of confidentiality), but I replicated the same search using the Timestamped JSI Web Corpus 2014-2021 (via SketchEngine) and interestingly came up with almost the same frequency order:

dogtooth*            532 examples                   
dog-tooth            115
dog’s tooth**      85
dogstooth            30
dogs-tooth            1

*I ignored capitalized forms because they were frequently proper names.
** Some of these were references to teeth belonging to actual dogs. So, I ignored examples of dog’s teeth, which were mostly about our furry friends’ gnashers and only counted the singular dog’s tooth which was more likely to be one of the other uses, although not always.

Then the question is how many of those variants do I include in my entry. If I throw them all in, will it just look messy and confusing? 

 

And do I trawl through all the corpus evidence to see which variants apply to which uses and try to show that?

Or do I just pick the most frequent forms? I can probably quite safely dump dogs-tooth, but where do I draw the line with the others? I don’t want people to look up, say dogstooth, because they found it in a text somewhere and think that it doesn’t exist or is “incorrect”, because it clearly isn’t. People are using it and presumably, understanding each other, so the descriptivist in me says that’s fine.

I won’t give away how I finally tackled the entry – back to confidentiality again – but next time you’re looking something up in a dictionary, take a look to see if there are spelling variants. Some poor lexicographer has probably agonized over them so they’re worth at least a glance.

 

Are you hyphen-hesitant?

At a recent webinar on EAP vocab, the topic of prefixes had come up and someone asked me the best way to teach students about when to hyphenate words with prefixes.

My first answer – which I wasn’t quick enough to give at the time – is I’d tell them to really not worry about it!

My second answer – and the one which I gave – is if they’re not sure, check in a dictionary. Although, to be honest, good luck with that …

Recently, I’ve been spending a lot of my time researching prefixed words and one of the things I’ve been checking is which are typically hyphenated and which are closed (no hyphen). And the results are really messy. It seems that a corpus search for almost any prefixed word will throw up a mix of both options.

Dictionaries: Dictionaries seem to vary a bit in their approach to showing prefixed words. Most though show the most frequent spelling as the headword, then the other alternative as a variant.

 

Although of course, different dictionaries use different corpora and will come up with different balances of hyphenated/closed examples. So, if you look up the same word in several dictionaries, don’t be surprised if you find different answers.

Some trends: Although it’s an area where there aren’t firm rules, there are a few tendencies:

Fixedness: Typically, newly created prefixed words tend to start off with a hyphen – perhaps to make the original root word clearer and the novelty less confusing. Then over time, as the combination becomes more familiar and fixed, the hyphen often gets dropped – think e-mail > email, on-line > online.

US vs UK: Also as a general rule of thumb, Brits tend to use more hyphens than Americans.

Double vowels: Where a prefix ends in a vowel and the root word starts with a vowel, hyphens are more common to avoid possible confusion over the pronunciation of the double vowel sound. So, co-opt is quite clearly /kəʊˈɒpt/ (two separate O sounds) whereas coopt not only looks a bit weird, but could potentially be pronounced /kuːpt/!

Looking at a few random corpus examples of prefixed words with double vowels, it’s clear that there are a mix of factors in play. The mispronouncability of the closed compound may be one – try antiageing without a hyphen. And some more familiar, fixed combinations, like preempt and preamble, are less likely to be hyphenated compared with more novel ones that err towards a hyphen – semiindependent is clearly a step too far for most people!

Note: All my corpus stats here are very rough and ready – this isn’t a careful academic analysis – but I think they probably represent the general trends. With this kind of thing, you also need to bear in mind all kinds of different factors, like the influence of style guides – so, for example, a major publisher or media group might decide at a certain point to shift from using e-mail to email and that could have quite an impact on corpus stats. As ever, language gets pushed and pulled by all kinds of sometimes mysterious forces!

Lexical layers 3: idiolect and finding your voice

Ever since I started getting interested in language, I've been intrigued by variation. For my undergraduate dissertation, I looked into differences between the way men and women use language. My MA dissertation looked at the effects of speaker age on the type of language we use – I found some surprising similarities between the youngest (teens) and oldest (70+) age groups.  Then when I recently dipped into the study of Forensic Linguistics, I read a lot about the concept of idiolect and how each individual's language use is shaped by a whole host of factors. This might include their gender and age, their social, cultural, educational and professional background, their social and political attitudes, where they grew up and have lived through their life and of course, their L1 and influences from other languages they speak.

Over time, we all acquire a slightly different set of vocabulary and we develop our own preferences about words and phrases we use and don't use, where, when and to whom. Some of those choices are quite conscious and considered, especially when it comes to sensitive areas like swearing and taboo or culturally sensitive topics. A lot of our choices though are largely unconscious. For example, do you mostly say thank you, thanks, ta or cheers? Or maybe something else altogether? Does it depend on the context or the person you're speaking to? Are there any of these that aren't within your personal idiolect? Personally, I'm not much of a cheers person. As a carefully-spoken, middle-class, middle-aged woman, it always feels a bit awkward to me. However, just occasionally, if I'm trying to sound a bit less posh and a bit more blokey, say speaking to a builder or a white van man, I sometimes slip in a "cheers, mate!" - always reverting slightly towards the London accent I grew up with. That might sound a bit patronizing, but in fact, it's a perfectly normal reaction and even has a technical name; accommodation (or more specifically convergence). But that's getting away from my point …

What does all this have to do with language learners? Let me give you a couple of examples. When I was at school my French teacher (an L1 French speaker) would sometimes exclaim in class when we were stuck on a question – "That's easy peasy, lemon squeezy!" We'd all laugh, in part, because it just sounded funny in her fairly strong French accent, but also because it's typically a childish expression you wouldn't expect a teacher to use.  Of course, she was well aware of what she was doing, it was a conscious choice to deflate a slightly tense atmosphere and it worked a treat.

My second example is of a student I taught several years ago. He was a late-teen, German L1 student visiting the UK. His English was good (B2+), he was keen to pick up spoken expressions and he would regularly use cheers to mean thank you. He (mostly) used it appropriately, but there was something about his accent and very precise articulation that made him sound like a very posh public school-boy clinking a champagne glass. I had to stifle a laugh every time he said it just because it sounded so incongruous.

Developing your own voice in a second language is tricky and no one wants to sound silly, but at the same time, we don't want learners to get overly self-conscious and never attempt to use new language. The good news, as I mentioned in my first post, is that a lot of basic, high-frequency language is, almost by definition, fairly safe and neutral which makes getting started relatively uncomplicated. It's as students progress and are increasingly exposed to a range of authentic language that more caution is needed. A quirky expression picked up from an online video that might be appropriate for the middle-aged male presenter in his local North London pub with his mates isn't necessarily going to sound right coming from a 20-something young woman on the other side of the world who drops it in with the American English she learnt at school!

That's not to say that we should be shielding students from all but the blandest, 'standard' English. Far from it! We should though be actively exploring the layers of meaning that the language they meet might hold. That doesn't necessarily have to involve explaining in detail all the possible subtle nuances of meaning and usage of a particular word, but at least flagging up language that students need to be careful about using. Emphasizing that they need to really get to know a word or expression, who typically uses it and when before they try it out for themselves. Synonyms can't just be substituted because they 'mean the same'. The language of journalism is intentionally colourful to draw in readers and may not work in a formal essay or a work email. The latest edgy youth slang will sound comical from a middle-aged mum. People generally use idioms for effect, to be playful or humorous, to play down the seriousness of a situation or to exaggerate it – writing "Hitler had a bee in his bonnet about the Jews" in a school essay (yes, that's a real example from a real student essay!) doesn't work on all kinds of levels!

In this series of posts, I've tried to move from the more obvious, broad-brush, outer layers of meaning, with basic distinctions of register and genre, down to the subtler, more difficult-to-define nuances of individual usage. These aren't issues that will crop up in every vocab set with every class, but if we don't explain this stuff where it is relevant, we're only telling half the story and we're seriously short-changing our students.