These days, any linguists who claimed to be prescriptivists would find themselves cast out of the linguistic community, and yet to a greater or less extent, we are all prescriptivists, in private at least, and those of us who teach a language to non-native students are necessarily prescriptivist – languages do have rules, and it is essential to at least try to inculcate them into the heads of learners (and anyone who complains about my splitting an infinitive in this sentence should read the original edition of Fowler’s Modern English Usage, published in 1926).

Descriptive linguists work with language as it is actually produced, and not like prescriptivists – who tell you what they would like it to be. The prescriptivist position is logically untenable – the first time a language is analysed, the analysis must be descriptive of the state of the language at that time: there is no previous analysis with which it can be compared, and so no “it must be”s or “it ought to”s are possible. So this has to have been the position that the first Greek grammarians found themselves in, and also the first Chinese, or Sanskrit, or any other linguists. In our own day, it is also what happens when linguists work with previously ‘uncontacted’ groups in places such as the Amazon.

Prescriptivists also do not seem to realise that their own use of whatever language they may speak is also vastly changed from that of their ancestors: they tend to accept their own usage as ‘correct’, even when it can be shown that their own grandparents must have had a different one (what was their grandparents’ attitude to change?). This is particularly easy to demonstrate with pronunciation, where they frequently point to the ‘mis-stressing’ of contribute and distribute on the first syllable, while themselves considering the stressing of the second syllable of vagary or quandary, which was normal 100 years ago, to be laughable.

These thoughts have come to me after reading Jack Windsor Lewis’s blog post on disputable pronunciations, where he describes /ˈerənəs/ for erroneous as an ‘undeniable mistake’. It could be that the eminent biologist who used this pronunciation (and I heard it too) believes that this is the correct pronunciation: after all homogeneous is often pronounced /həˈmɒdʒənəs/, even though there is a word of a different meaning spelled homogenous, and while this pronunciation is not yet given by EPD, ODP or LPD, the Oxford BBC Guide to Pronunciation does have it, with no condemnatory comment: “Less commonly also pronounced huh-moj-uh-nuhss”. In my recent experience, the only thing wrong with this statement is the two words “less commonly”, because I now seem to hear it far more often than the ‘regular’ pronunciation /hɒməˈdʒiːnɪəs/. Are these people also making “an undeniable mistake” in Jack’s words, or as descriptivists, must we now say that it is a part of the language? The editors of all but the BBC dictionary seem to be saying that it is not acceptable – they must have heard it, and yet even John Wells, who marks all sorts of pronunciations as “not acceptable” (and this is another example of a great descriptivist being prescriptive), denies it an entry. If Olausson and Sangster are correct, on the other hand, then /ˈerənəs/ must be considered a “less common” pronunciation of erroneous.

I have heard another of our greatest linguists telling an audience that it is impossible for a native speaker to make a mistake. This must surely be wrong – mistakes are made by accident, and we have only to consider spoonerisms, which are inadvertent, to realise that we all make mistakes in our speech every day of our lives.

Prescriptivists are clearly losing a fighting battle – as I once said – in trying to stop all change that they happen not to like, but descriptivists too will correct what they consider to be mistakes. The problem is: when does a mistake happen so often that it ceases to be a mistake, and becomes an acceptable alternative, or even drives out the older version altogether, as has happened with vagary and quandary ( and /bælˈkəʊni/ a hundred years earlier)?

The Germanic group of Indo-European languages split into three sub-groups – East, West and North Germanic. East Germanic, now extinct, included Gothic, in which the earliest  example of a Germanic language in writing – the extracts from the Bible translated by Ulfilas – still exists. This manuscript dates to the 4th century, and is housed at Uppsala in Sweden (having been looted from Prague during the Thirty Years War).

West Germanic gave rise to present-day German, Dutch, English, Frisian and Afrikaans.

North Germanic, which comprises the Scandinavian languages, itself divided into two: West and East. West Scandinavian is now represented by Icelandic and Norwegian (and Faroese) while East Scandinavian is Danish and Swedish. These are the historical divisions, but there are various features of the languages which cross these boundaries in different ways.

The most obvious is that Swedish and Norwegian share ‘tonelag’ – a feature unusual in European languages, of tonal distinctions like those in Chinese. Pairs of two-syllable words may be distinguished solely by whether they are spoken on Tone 1 or Tone 2. However, in dialects of Northern Norway, this feature has disappeared, leaving pairs such as “bønder’ (“farmers”) and ‘bønner’ (“beans”) homophonous. I don’t know if the same has happened in the Swedish spoken in the north of that country – perhaps someone can enlighten me? To complicate matters still further, the tones may vary from one dialect to another, just like vowels and consonants.

Neither Danish nor Icelandic have ‘tonelag’. Danish, however, does have a reflex of it, called ‘stød’, which is a glottal stop inserted in one member of the pair. Just as the spelling of Norwegian and Swedish gives no indication of the tone pattern on their bisyllables, so Danish does not reliably indicate in its spelling the presence or absence of ‘stød’ – and an added complication is that sometimes the glottal stop, which comes in the coda of the syllable, precedes a final consonant, and at other times follows it. The dialects of the extreme south of Jutland have lost the stød – which puts them in a similar situation to the dialects of northern Norway. This is fortunate for foreign learners, who can just ignore both tonelag and stød without causing too much confusion in their listeners!

On top of this, Norwegian Bokmål often looks like Danish, while Nynorsk can resemble Swedish. In Danish, the word for “week”, for instance, is ‘uge’. In Bokmål it is ‘uke’. In Nynorsk it is ‘veke’, and in Swedish ‘vecka’. In Icelandic, ‘vika’. Here Bokmål and Danish go together. Icelandic – which ‘should’ be with Norwegian, agrees with Nynorsk, but both are close to Swedish.

The isoglosses are difficult to sort out across the Norse area. Icelandic has retained a lot of features of Old Norse, and is more different from the other three than these are amongst themselves. The speakers of  Danish, Swedish and Norwegian get along very well without the need for translators, and it always surprises me that fiction for adults written in one of the three may be translated into the other two.

The tragic and shocking news from Norway in the last days has brought the name of this formerly insignificant island in a lake not far from Oslo to the world’s attention.

The name means “the outer island”, illustrating the Scandinavian feature of placing the definite article, in this case feminine, after the noun as a suffix. It also has the typical Scandinavian (Norwegian and Danish, but not Swedish or Icelandic) letter <ø> which represents a front rounded vowel. The British media are mostly replacing it with <oe>, treating it (quite reasonably) in the same way that they would the German (and Swedish and Icelandic) <ö>, or ‘o umlaut’.

The pronunciation in Norwegian is [ˈʉtœya]. The [œ] is rather different from the French vowel represented by the same IPA character, the French vowel having ‘close’ rounding, while the Norwegian one has ‘open’ rounding – the lips are much further apart.

This is obviously causing problems to English-language broadcasters, not surprisingly, but the nearest one could come would be something like /ˈuːtˌəjə/.

The word “utøy” (without the definite article) is ambiguous – splitting it into ‘u’ and ‘tøy’, it means “vermin”, and the pronunciation is distinguished by the /t/ being strongly aspirated in “vermin”: [ˈʉtʰœy], but not in “outer island”. “Vermin” is also neuter, and so in the definite form is ‘utøyet’ rather than ‘utøya’.

I’ve just come across this, entitled “Learning Pronunciation with Dictionary IPA” and was appalled. Macquarie University is quite rightly acclaimed as the home of the standard Dictionary for Australian English, which you would think would entitle users of the language pages of its website to expect a high standard of scholarship. Yet this document appears to have been put together by non-native speakers of English, who have not had their English proof-read by a native speaker, and who also have a very defective knowledge of phonetics and what we might call the “phonetics world”.

At the head of the document we read “The Dictionary International Phonetics Alphabets (IPA) is a set of phonetic transcriptions …” Since when has “Alphabet” been plural in this context?

Then we come to “The Basics”:

“Syllable is … a unit of spoken language and consists of one or more vowel sounds alone or of a syllabic consonant alone, e.g. the word sy/lla/bles/ consists of three units of sound.” So how do the consonants fit into this, and is the final syllable ( in the terms of the document) the ‘l’ or the ‘e’?

The final sentence defining “Vowel” states: “Sometimes vowels are referred as ‘voiced’ sounds.” This could simply be a typo, with ‘to’ (between ‘referred’ and ‘as’) omitted by mistake, but under “Consonant” we have “Sometimes consonant is referred as ‘unvoiced’ sounds.” Poor English, and just plain phonetically wrong.

So, is a syllable one sound, or is it made up of more than one sound? The writer does not tell us, and continues to confuse us to the end (see below as well).

The next paragraph answers the question “What is the International Phonetic Alphabets?” So, once again, an unexpected plural. This is beginning to sound like the meerkats in the British TV advert for a price comparison website (for readers outside Britain, change ‘meerkat’ into ‘market’ for an explanation of this).

“How does IPA work? The tables below show how each symbol sounds like …” Shouldn’t that be ‘what’, not ‘how’?

“Put the sounds together … For example, you have the word ‘examine’. First, divide the word up into syllables (single sound), e.g. the word, e/xa/mine, (so three syllables). Try to pronounce the word. Next, Look up the word in the dictionary and you will find the IPA of the word /examine/, the IPA is /ɪ g ˈz æ m ɪ n/ (British pronunciation). Then, check where the stress is. Is the primary stress in /e/ or /xa/ or /mine/? It’s /ˈxa/. So you sound louder, longer and done in a higher pitch of voice on the sound /ˈzæ/”.

There is total confusion between spelling and transcription, where both are enclosed in /…/. How are learners to understand that the orthographic <x> is split between two syllables when they are being told that the spelling has <xa> as the second syllable, but the transcription has /z æ/, and what is one to make of the statement “you sound louder, longer and done on a higher pitch”?

I know the Internet is full of rubbish, but for this to appear under the ‘imprint’ of a respected university in a country where the principal first language is English beggars belief.

Rant over.

The addition of a comprehensive pronunciation history to all the entries in a dictionary would be wonderful – and comments to my last post confirm my opinion. However, there are just a few slight problems.

Jack Windsor Lewis has pointed out that the OED gives a full list of spelling variants through the ages for each word in the etymology section, and implies that we can learn a lot about the word’s pronunciation from that, but while spelling several hundred years ago was at the whim of the writer, and so was likely to reflect that writer’s pronunciation, the fact that several spellings were often used contemporaneously with each other probably means that the pronunciation also varied at the same time.

Johnson, in the exchange with Boswell that I quoted last time, says “whose pronuncation should we follow?” Does a historical (or an historical – thank you Martin) pronouncing dictionary restrict itself to one dialect/accent, and if so, which one? or does it cover the range of dialects/accents at each period?

As if this was not a thorny enough problem, the next is even harder. How do you represent the pronunciations? All the current pronouncing dictionaries use a phonological transcription, and have done since Daniel Jones’ first edition of the EPD in 1917. The OED, in its original form, used a transcription that was much more complicated. There is a very good explanation of the history of this transcription, and also its interpretation, by Professor Mike MacMahon in the Transactions of the Philological Society, 1985, pages 72-112.

A phonological transcription is all very well, but the symbols used in the EPD, for instance, have barely changed since 1917. A.C Gimson, in his 1967 edition (the 13th), changed /ou/ to /əu/, and in the 14th edition, further amended /əu/ and /au/ to /əʊ/ and /aʊ/. He also updated the vowel charts in the introduction to reflect changing qualities. This helps to cover a single changing dialect/accent, but without a large number of explanatory notes, it could not explain the full variety of English pronunciations, whether diachronically or synchronically. A truly phonetic transcription would need to be used, which might result in entries looking something like this (which I hope is approximately accurate):

ˈhlaːvwɛərd > ˈlaːvərd > ˈlɔːvərd > ˈlɔːərd > ˈlɔːrd >ˈlɔːəd > ˈlɔəd > ˈlɔːd

(hlafweard > lord).

And this needs to be repeated for every word, from its first appearance in the language. As Martin Ball has asked, are there any takers for the job?