Numbers in 5000 Languages -website review

Count to 10 in almost every known language. This simple idea for a website is the perfect premise for a broad overview of the vast array of human languages. By giving a little taster of each, it is possible to easily see the relationships between members of language families, and see how languages have developed and varied across the globe and through history.
Take the Indo-European family of languages. They arose in a broad arc of the globe from Iceland, across Europe (including English) through to northern India. This vast family is believed to have its origins in the Caspian Sea area several thousand years ago and spread through migration and contact to Europe and India and splintered into many of the world's major languages. This website easily shows this 'genetic' link by listing numbers 1-10. A few examples:

English: one, two, three, four, five, six, seven, eight, nine, ten
Welsh: un, dau, tri, pedwar, pump, chwech, saith, wyth, naw, deg
Russian: odín, dva, tri, chety're, pyat', shest', sem', vósem', dévyat', désyat'
Bengali: æk, dui, tin, car, pãc, choy, sat, at, noy, dosh

The notes on the site throw up some interesting insights into other languages:
  • Numbers in Yoruba (Africa) are built on substraction, so 46 is sixty-less-ten-less-four.
  • In Kumbundu (Africa) the number 7 is a euphemism: to say 7 directly is taboo, so they use a word derived from '6+2' (I guess this is no more weird than numbering the 13th house 12A, or talking about the 'Scottish play').
  • Yumbri (Thailand) claims to have no numbers, only little and much.
  • There are a few base 5 systems, but there are some that are base 4 (so 7 is '4 and 3'), base 8 and binary.

And if the existing 5000 languages are not enough, there is also a slightly nerdy section on imaginary and sci-fi languages, as well as a bewilderingly optimistic array of artificial languages.

Strangely sign languages are omitted. Although they all (as far as I know) use the fingers for counting, each sign language has different ways of doing it. British Sign Language numbers and American SL numbers have only 1, 2, 4 and 5 in common:



Interestingly, BSL for eleven and twelve are (like English and French) irregular, i.e. they don't follow the pattern for the rest of the teens.

Link for Numbers in 5000 Languages

Bluffer's Guide to Passing Exams

One of my hobbies is doing multiple choice tests in subjects I know nothing about and seeing how high I can score. Today I got 45% in an online test of Russian (, putting me at A2 level (i.e. one above beginner) and my language was described as follows:
"You will be able to understand and use sentences and frequently used expressions related to areas of everyday speech. (e.g. very basic personal and family information, shopping, local geography). You will be able to participate in a conversation concerning personal interests."
No I can't. I don't know any Russian - I can't even read the alphabet. Yet I managed to score correctly in nearly half the questions, a full 20% higher than what would be expected by chance.
Also today I did an advanced Dutch test, and was placed in the 3rd level (pre-intermediate)
Two years ago I did a test for doctors applying for immigration to Australia (Australian Medical Council - sadly the test has been removed now), I got 23/50, two short of the pass mark. Again I know nothing about medicine - I couldn't understand the language in most of the questions, let alone the answers. Added to this, there were 5 choices, giving an expected random hit of just 10/50.
How is this done? Well, its all to do with the Multiple Choice Test.
If somebody asks me a question in Russian, I am certain to give the wrong answer, because my knowledge of Russian is zero. But if I am given me four choices, then I have at least a one-in-four chance of getting it right. My success in the test depends, not on my skill in producing a complex foreign tongue, but simply my skill in picking a, b, c or d. People think that if there are four choices, then a person with zero knowledge must have only a one-in-four chance of getting it right. This is only true if each option is equally plausible. It only takes one option to seem unlikely to increase your chances to 33%; two unlikely answers and we're up to fifty-fifty.
So, what are the secrets? Here is my easy guide to bluffing to multiple choice test:
Choose the answer that is most like the others. Here is an example from a real test:
a. ground
b. arena
c. pound
d. stadium

A weakness of Multiple Choice is that writers naturally write distractors (wrong options) that are similar to the right answer. Answer a above is the only option that has something in common with all the other three (why would pound be an option when the only thing it has in common with the others is that it rhymes with ground?). Knowing this, it is often possible to triangulate the correct answer, or at least improve your odds.

Here is set of options from the Russian test:

a. входить - войдите
b. входи - войти
c. войти - входите
d. войди - входить

It looks tough at first sight, but I chose c because it is the only one where both words also occur in the other options. This is the correct answer.
Often there is one option that is obviously different from the others, or a nonsense answer. At other time two anwers look very similar: the correct answer is likely to be one of these two.
Another trap that question setters sometime fall into is making the correct answer longer in length than the distractors - this accounted for many of the correct answers in the Australian doctor's test.

Multiple choice is one of the most common formats for testing, often in important and high-stakes exams (no doubt because it's administratively easy to mark), but the margin between chance and skilled performance is sometimes so narrow that it's quite easy for these tests to have compromised validity.

Sign here

My boyfriend is deaf. When I met him 5 ½ years ago, he was only the second deaf person I had ever properly met who didn't have any spoken English. Although we could communicate satisfactorily through writing, it immediately became clear to me that I would have to learn sign language.
First, I had to overcome some preconceptions and misapprehensions:

Sign languages are languages comparable to any other human natural language
Sign language uses mainly the same language areas of the brain that are used in spoken language, rather than the motor areas that are used in, say, juggling (I experience an interesting phenomenon when John 'shouts' – makes energetic signs – in BSL: my ears hurt as if someone is physically talking loudly; my friend Les, while struggling to remember the sign for 'rain' blurted out 'la pluie', interchanging a signed 'foreign' word for a spoken one). BSL has all the features that define language, including arbitrary form-meaning relationships, a finite number of components that recombine to make an infinite number of utterances. It has phonology (sometimes called cheirology), grammar (actually very distinct from English), it has metaphor and idiom, puns, poetry and rhyme (using similarity in hand shape rather than similarity in sound), slang, accents, dialects, formality and informality etc, etc.
One major area in which BSL is lacking is that it has no written form and therefore no body of written literature.

It is a natural language
This means, like all the other mother tongues on Earth, it is another manifestation of the universal human language instinct and evolved spontaneously as a result of a community of people trying to communicate. It was not invented by missionaries, nurses or teachers. An artificial British sign language was invented in the mid 20th century called the Paget-Gorman system, which, as far as I know, has largely disappeared under the resurgence of BSL, going the way of all well-intentioned artificial languages (which are doomed to failure for the same reasons that genetic engineering cannot match the results of natural selection).

BSL is the 3rd most used native language of the British Isles

It has 50-70,000 deaf users, added to a considerable (but undetermined) number of hearing bilinguals and learners. Number 2 is Welsh.

It has a long history

The first record of sign language between deaf people in Britain was a wedding in Leicester in the 16th century where the vows were signed. Daniel Defoe published a fingerspelling alphabet (pictured above). It's reasonable to assume that anywhere where deaf people lived together that signing of some kind developed (as happened when Nicaragua opened its first deaf school in the 1980s, when, for the only time in history, researchers witnessed the birth of a new language). BSL was suppressed in the early 20th century by educationalists who believed that signing would prevent a child from learning speech and lip reading, with the monstrous result that many deaf children grew up without a first language, without a medium to fully express themselves.

It is structurally unrelated to English and has its own grammar

  • It has classifiers (like Chinese), where nouns are grouped according to their characteristics (e.g. animal, wide & flat, tall & thin, round etc), and has a far wider range of pronouns than English (e.g. an inanimate singular noun can only be represented in English by it, but in BSL there are a whole range of inanimate pronouns depending on their characteristics).

  • Verbs can be modified by their subject or object.

  • Word order is flexible (unlike strict SUBJECT-VERB-OBJECT in English), and the strongest organising principle is (in common with many languages, including Japanese) TOPIC-COMMENT (e.g. exam, Zoe's passed or Mary, she's pregnant).

  • BSL, (in common with about half of native American, African and Papuan languages) distinguishes between inalienable (permanent) possession and alienable (temporary) possession, so the my in my brother takes a different form to the my in my cup of coffee.

Although it has many features that are vastly different to English, BSL, with its users being thinly scattered throughout an English-speaking nation, is heavily influenced by the dominant tongue. The fingerspelling system represents English letters when signing names of people and places, and are incorporated into many signs (many family relationships and days of the week use the initial letters of the English words; do and if can be spelled out as in English). Lip movement is an important component of a sign (it's not all just in the hands), and often silently follows the lip pattern of the equivalent English word; the sign for husband and wife both indicate the ring, the difference showing in the lips movement. Sign language can also follow closely English word order and grammar in a form of sign language called Sign Supported English (SSE). In its most 'English' form, SSE is used to teach profoundly deaf children English and literacy, using one sign = one English word and even sometimes including the English morphological endings, like -s, -ed, -ing; in its common form, it is a preferred form of signing for many users (depending on their education or upbringing), which has many elements of BSL, but with a more English pattern (e.g. SSE would sign the dog jumped over the fence DOG JUMP OVER FENCE; BSL would sign DOG then make the classifier for animal in the right hand and move it over the left hand representing FENCE). There is a continuum between BSL and SSE, different users moving between one or the other in different situations.

Sign Languages have some features not commonly found in spoken tongues

  • Multiple signs can be produced simultaneously, i.e. one in each hand + face. They were surprised to meet each other on the stairs can only be expressed in English as a string, one word at a time. In sign language these 10 words can be shown (as it actually happened in real life) as one event in a single sign. This is how (even though hands move slower than tongues) sign translations take the same amount of time as speech.
  • Use is made of space to accurately show relationships between things. Try explaining the offside rule in English without using your hands or a diagram: it's sooo much easier and more accurate in sign language.
  • Unlike spoken words, many signs have a physical or metaphorical similarity to the things they represent, making sign vocabulary easy to learn, improvisation and creativity possible, and a more subtle and poetic representation of many ideas.

Sign Language is not international
While BSL and Auslan (in Australia) and New Zealand sign language are related and mutually intelligable, Irish Sign language is not. American Sign Language is (because of the quirks of deaf education history) related to French sign language, and is not comprehensible to either Irish or British signers. So, while the patterns of distribution are different, sign languages around the world vary as they do for spoken languages, and for the the same reasons. Languages are codes of cooperation maintained through interconnected webs of communicators: where connections are rare, the codes will inevitably diverge.
There is an international sign language, Gestuno, used in international deaf conferences. I've never met anyone who uses it, but I suspect either it's heavily biased (like Esperanto) towards a small group of natural sign languages, or else it's so general it pleases nobody.

Russell Brand and me

I have something in common with Russell Brand. No, I have never been called 'Byronic' (not to my face anyway), my comedy series only happened in my mind, and I have never been the victim of a media bandwagon (except for that time when my doppelganger was a wanted IRA terrorist - true!). No, Russell Brand used to teach English.

According to EL Gazette:
"He had been an EFL teacher on London's Oxford Street for a year. Describing the quality of his teaching as less than good, he admitted that he was the 'cool, popular teacher' who, following a time-honoured Oxford Street private EFL sector practice, would occasionally take his students for lessons in a nearby park on sunny days. When Brand started passing a joint among his students, however, he was betrayed by 'some evil student who actually wanted to learn English!' Brand recalled desperately preparing his class by getting them to collectively agree an alibi for him. That stratagem failed however, because in Brand's own words, 'I was such a shit teacher that none of them understood me.'"

My favourite word

English: unbeknownst

BSL: recently

What do foreign learners find difficult about English? Part 2

One the earliest and most basic things you need to be able to do in a language is to ask a question. Otherwise you can't get what you want. In French you can add est-ce que to the front of the statement. In Japanese you simply add the particle -ka on to the end of the last word in the sentence. There – you have instantly learned how to ask a question in Japanese.
It's not so easy in English. Questions are hellish for a beginner.
It all starts off quite easily with the subject-verb swap common in Western European languages:
She is Irish > Is she Irish?(words that change position are underlined)
But that only works for certain verb such as be, can, could, will, would.
It gets a bit more complicated with multi-part verbs. You have to make sure you do the swap with right part:
She is working in London > Is she working in London?
She has been working in London. > Has she been working in London?
So far, not bad. But applying the same rules:
She works in London. > Works she in London? No, we (in contemporary English) have to say: DOES she work in London?
So how would you explain the rules that creates this transformation (rules, by the way, that you faultlessly follow every day)?
She works in London > DOES she work in London? (Words that change form are in bold; words added in CAPITALS)

Here are the rules:

  1. Take the verb do
  2. Look at the tense of the main verb (works) and put do in the same tense > do
  3. Make it agree with the subject (she)> does
  4. Place it before the subject > does she
  5. Take the main verb (works) and change it to the base form >work

The person you're asking has left the building.

Similarly in the past tense:
She worked in London. > DID she work in London?

Not only that, but the form of the question can depend on what missing information is being asked for:
Sarah gave the tickets to Steve. > WHO gave the tickets to Steve?
But we can't (in a genuine question, rather than for showing surprise) do the parallel expression: Sarah gave the tickets to WHO? We have to say: WHO DID Sarah give the tickets to?
(And that's before we even consider when it is 'correct' to use whom...)

Another difficulty is how many words to move to the front.
She works in London three times a week. > How many does she work times a week? NO
>How many times does she work in London? NO

>How many times a week in London does she work? NO
HOW MANY times a week DOES she work in London? CORRECT!

I hasten to add that this is not how I teach questions to my students - I have tried to give the learner's-eye view and emphasise the bewildering aspect of question formation. If you go 'one level up' there are actually quite simple and beautifully precise rules that enable us to generate questions. The problem for the learner is that they have to have quite an advanced knowledge of English syntax to be able to get a clear view of it all.
It's a lot to ask for, when you want something.

Killer in the class

One of my students told me today how he killed somebody.

We were practicing the simple past tense with the question When did you last lose something? And he told me how, in his country, somebody stole his wallet and he pursued him in his car.
'He died when I hit him'
'So he ran out in front of you and you couldn't stop in time?'
'No, I chased until I hit him'
' did you feel?
'I was scared'
'I'm not surprised. I would feel bad about killing somebody.'
'I was scared about the police'
'Oh. What did the police say?
'They said it was OK. Nothing happened'. trace of regret or emotion. And he had seemed like such a nice boy.

It's amazing what students will reveal in class. Generally teachers encourage bringing personal experience into language learning, but at the same time we hope the students keep things bland enough to avoid a diplomatic or emotional crisis. In one of the first classes I ever taught, I went round the class asking What do you like? What do you dislike? Expecting answers like ice cream, opera, noodles, I came to the Palestinian teenager. He stood up and shouted 'I hate the Jews for taking the land from our people!'. Luckily the only other people in the class were two Chinese people whose English was too weak to understand what was going on.

What do foreign learners find difficult about English? Part 1

The learner has to tune in to a new system for making sense of sound
English has about 44 phonemes – individual sounds which form the smallest meaningful unit in a language ( This is, although by no means huge, a larger sound palette than many other languages (the number of phonemes in Spanish is in the mid 20s). The problems when mapping these onto the Roman alphabet of just 26 letters is obvious. There is nothing about h + c on their own that would lead you to think that, combined, they make the first and last sound of church? That phoneme is really a combination of sh and t, so it would more sensibly be written tsh; I went to tshurtsh. That's why cat shit sounds like catch it. There is no way of distinguishing in the alphabet the two phonemes represented by th in thigh and thy. The various ways of pronouncing gh is famous. This vestigial spelling bears witness to an extinct phoneme in standard English – the glottal fricative found in Scots loch (lough in Northern Ireland, or Van Gogh in Dutch).
English has a large number of vowel phonemes – about 20 depending on how you count – compared with 5 in Spanish. This enables us to have a large number of one-syllable words:
After 5 words, Spanish would have had to add a second syllable – a reason why Spanish words are often longer than English ones. While this makes English very concise, it has the disadvantage of placing an extra burden on the learner. The vowel palette is a finite space, and in order to have more vowels, the distinctions between them must be finer. The vowels in fit and feet form one the most productive contrasts (hundreds of words differ only in respect of this vowel), yet the physical difference is tiny – a minuscule difference in tongue position and (sometimes) a slight difference in length – and the auditory distinction approaches the limits of human perception. As a result many foreign learners struggle to appreciate the difference between such words as
and are constantly in danger of saying such things as a shit of paper or a piss of paper. Even many advanced learners will swear that the two sounds sound identical to them. An English person, meanwhile, (when listening to a native speaker with a familiar accent) never confuses these sounds, just making use of the similarity in puns like life's a beach.
The problem is, if you don't need to hear the difference in your own language, it's hard to hear the difference in a new one. For example the words hit, hat, hot all start with the same sound don't they? No they don't. Say them – each h has a very different lip shape and sound. The a in add is different from the a in mad, but it's not important in English so we don't waste our processing capacity by ever noticing it. In other languages the two a's might form a meaningful contrast. In Spanish the r in pero (but) is different (longer and more rolled) to the r in perro (dog). To us they sound the same, or just two different ways of saying r. To a Spaniard they are separate phonemes. A Japanese person famously groups English l and r together, yet we group together the first and last sound in little as the same l sound (in fact the mouth is different, and the sound is different, but we just don't notice it) while other languages hear them as differently as we do d and t or s and z. In fact there is often more variation within one phoneme than there is between different phonemes. It all depends where the boundary between two phonemes are. Experiments have been done where a phoneme is recorded and repeated electronically, each time moving closer to a neighbouring phoneme, say hid moving towards heed. At no point in the middle will a native speaker think that it sounds halfway between hid and heed; it will sound at most like a slightly weird hid before suddenly and completely jumping to heed. This is the gestalt effect in perception, like the silhouette of the vase and faces – you completely flip between the two categories and you never see both or neither.
The disparity between the perceived and the actual (acoustic) differences between phonemes is demonstrated when we talk on the telephone. In a normal conversation we can say
Peter and Paul have bought a boat.
And we would never hear it as
Beter and Baul have pought a poat.
The first sentence is the interpretation that make more sense, so we hear what we expect to hear (top-down processing).
But when it comes to spelling a postcode or name, there is no context, so have to rely much more on interpreting the actual acoustic signal (bottom-up processing) ,so BP on the telephone can be impossible to distinguish from PB, BB or PP
It's only at this point that we realise B and P are, in the slightly impoverished medium of the phone, very similar sounds, and it's at this point we have to flesh out the sounds with
B for Bertram
P for Parker Knoll
It's not just that different phonemes can sound similar, it's also that the same phonemes often sound different. Writing has over simplified word structure: writing is made up of 26 letters that can be recombined to make every word, and, while every word is different, the letters themselves remain the same. In actual speech, although we perceive 44 phonemes remaining intact as we recombine them, they actually vary enormously depending on the neighbours. More on this another time, though.
So, in short, a learner listening to any new language has to learn that, on the one hand, there are subtle, tiny differences between sound categories in the new language where an inch is a good as a mile, and on the other hand, large differences between sounds that native speakers seem to ignore. That's why voice recognition software has (until recently) been so unsuccessful. And that's also the challenge for a learner of any new language.

Apostrophe rules OK

"Now children will go around Birmigham and see utter chaos" John Richards, founder of the Apostrophe Protection Society

Birmingham City Council are removing its apostrophes from road signs and street names such as Kings Norton and Druids Heath (Metro, 30/1/09). The Apostrophe Protection Society condemned this as 'retrograde and dumbing down. It's a bad example to children and teachers'.
It's sad that nearly every time English is in the papers it is at this level of pettiness, railing against the 'misapplication' of a tiny range of popular prescriptive rules. When I started this blog, I intended not to get drawn into this kind of thing (in fact it's hard to find a language website that isn't obsessed by this kind of thing), but this is a very good example of self appointed 'experts' wading in and over applying their little and outdated knowledge.
The prescriptivists are mistaken for two main reasons (borrowing from David Crystal):

  1. Language varies

  2. The 'grammarians' get the rules wrong

Language varies enormously across time, across space and between and within individuals. There are in the region of 5,000 living languages which break down into an estimated 20,000 dialects, added to an untold number of dead languages and dialects. Each individual human being speaks in a slightly different way compared to their neighbours (an idiolect). An individual's language will change throughout one lifetime (compare David Attenborough's accent in the 1950s to the way he speaks now, or the Queen's). We vary our language in subtle but systematic ways according to context. Language has always changed and always will do – not just an observed fact but a necessity. It is an ecosystem of millions of variations in a permanent and enriching state of flux. If you think English needs to be preserved, remember it was created, not by grammarians imposing their rules, but largely unconsciously by generation after generation of uneducated and illiterate peasants just saying whatever they felt like saying. So how can any one language feature claim to be permanently correct?

People get the rules wrong. A little knowledge is a dangerous thing (and there is a lot of little knowledge about language), and so often an observed rule (the taxonomy that helps us find patterns in natural language, or helps students learn a foreign language) becomes a prescribed rule and the skilled native speaker is told the way they speak naturally isn't quite 'correct'. The problem is that these explicit rules are rarely completely right. An example. In many standard grammars of the 20th century, shall was considered the first person form of will; I shall and you will were correct, I will and you shall, according to self-appointed experts, were 'grammatically' wrong. Modern linguists now have a different explanation. Will and shall are separate verbs, and both can be used to talk about the future; the difference with shall is that it implies a stronger involvement of the speaker in the predicted events. So, I shall leave tomorrow means 'I will see to it that I leave'. You shall leave tomorrow means 'I will see to it that you leave'. We can see that the involvement of speaker implied by shall is quite natural in the first person I – we normally control our own destinies - but when we switch to you or they, we are imposing our will on other people and this is less acceptable. You will go to the ball can be a neutral prediction, whereas you shall go to the ball involves 'I' as the speaker as master of your fate. The difference between will and shall, then, is much more one of semantics than of grammar. As a result of this artificial rule I shall, you will, a dissonance was created between what people were told was correct and what came spontaneously to them, thus distorting the natural and useful distinction between will and shall. Nowadays, people aren't really sure of when to use them, and often shall is just an old-fashioned or formal version of will. A perfectly good pair of words that everyone was using quite happily until grammarians with half-baked ideas came along and upset the natural balance of the ecosystem. This is just one example of a half-understood rule replacing or distorting the unconscious, and (nearly always) far more subtle, sophisticated and systematic use of English that a native speaker - whatever their educational achievement - effortlessly commands. No one (not even Chomsky) is able to state all the rules of their language (certainly not enough to dictate to others), yet every native-speaker can employ them all with ease.

For a wonderfully detailed and succinct argument against prescriptivism and 'the rules', I would encourage you to read Stephen Fry's blog:

Apostrophes in names.

Back to apostrophes. Should the Birmingham road signs have apostrophes or not?

There is a standard rule that NOUN+s marks a plural and NOUN+'s a possessive. Fine - that's a very reliable rule for standard English. The Apostrophe Protection Society think therefore names (including road signs) that indicate a possessive should have an apostrophe, e.g. Starbuck's, not Starbucks, King's Norton, not Kings Norton. What these apostrophe-huggers don't understand is other principles influence English apart from their one precious little rule.

Names behave differently to common nouns:

  • Names can be spelt 'wrong': Wilde, Whyte, Hogg, Kwik Save. In fact it's very usual for names to be spelt differently from their common equivalents. Like the capital letter, it serves a useful function to distinguish a name from an ordinary noun.

  • Singular things can have 'plural' names: Waters, Richards, Hills

  • Names can have possessive origin, without needing an apostrophe: Davidson, not David's Son

  • Names are arbitrary - they don't have to 'mean what they say': New Road is one of the oldest roads in many towns. Mrs Goodman could be an awful woman. You absolutely don't need to know that Napoli is Greek for new town for it to work as a name.

  • Names can exist completely outside the English language, as in foreign names, Prince's symbol name, or company logos.

English grammar is weak insude names, and this is as it should be. It's not because of 'slovenliness', but for very sound reasons that reach beyond the surface conventions of language. The only function of a name is to provide a direct reference to a particular object - a simple label or pointer - and as such it doesn't matter what form it takes. You can give your child or cat whatever name you like and spell it however you want - it doesn't need to make any sense at all beyond simply connecting itself to a single object, and with that connection, it's purpose is fulfilled. In the terminology of semantics, a name has reference, but no sense. When I say cat on its own, it has sense - I have a generic model of what a typical cat is - but I am not referring to any single cat in the real world. But when I add grammar - my cat, Bill's cat, that cat over there, the cat that lives next door - I am able to zoom in and identify one single unique instance of a 'cat': this is reference. Or, I don't need to use the word cat at all and just give it a name, such as Mimi, which has reference (it identifies a unique individual) but no sense (Mimi on its own is meaningless; it only means something if you already know it to be the name of my cat; the word cat on the other hand has sense for every speaker of English). A name, as a direct label of one unique object, does not require any internal grammar because it is does not depend on grammar for its comprehension. This is not about language rules, but the logical and universal principles that exist independently of any language. By artificially enforcing an apostrophe rule in names, the grammatical do-gooders are fighting a losing battle against far greater forces. Grammar is necessary in order for words with sense (abstract, general, 'dictionary' meaning) to take on reference (labelling a unique object); grammar is superfluous for a name, because it already possesses intrinsic reference. In other words, grammar is necessary to helps us understand what people are talking about, but with names the job is already done for us.

So why do some names have apostrophes, and some not?

Although we have established that names are a low-gravity zone for grammar, we can detect some principled forces at work in the application of apostrophes in names.

There are 'degrees of possession' in names. This is related to how much internal meaning we see in the names.

  1. Bill's Cafe is very much possessive – we are invited to picture Bill standing proudly in an apron next to his cafe. This name is clearly descriptive - a cafe owned by Bill. It has a meaning (sense) as well as being a name. Therefore the name has an apostrophe to aid this description.

  2. Sainsbury's: again the apostrophe is used in the name. While Lord Sainsbury may seem a more distant figure than Bill, we still sense (or the company wants you to think) that this is a shop founded by someone called Sainsbury.

  3. Starbucks: here, no apostrophe - a name that APS thinks is wrong. But the cafe does not belong to Starbuck (being a fictional character in Moby Dick) - it is just a name - so we feel less obligation to use the possessive. Same with Boots.

  4. Tesco: here the 'official' name is Tesco, but it is commonly referred to as Tesco's (Marks and Spencer is similar). We don't mean to imply that it is a shop owned by Tesco in the way that Bill does, but we do like to add an 's to the end.

  5. ASDA: here both the official and common names are without s - I'm not aware of it being called ASDA's. Another example is Holland and Barrett - clearly names of people, but no s. We see them as wholly proper names.

  6. There is an element of 'free variation' (where no principles seem to apply), as you would expect with the arbitrary nature of names, and their weak demand for grammar. So the index of the London A-Z reveals St James Road, St James' Road and St James's Road. I have a sneaking suspicion that the older a name is, the more it loses its grammar (as is the case with compound nouns moving from 2 words door handle, to hypenated hat-trick, to one fused word cupboard). Maybe someone could investigate.

But if Starbucks and Boots are not possessive or plural, then why is s added to the end? I would suggest that there is a strong tradition for shops and cafes to end with s. Maybe it evokes a cosy image of Bill, the proud proprietor, a distant echo of the times of owner-shopkeepers, hints of a possession that no longer fully applies; maybe it's just habit; maybe it just rolls off the tongue easier. Starbucks and Boots sound cosier than Starbuck or Boot, more reflective of a welcoming establishment, while Starbuck's and Boot's assert an ownership (through the overly grammatical - and common -apostrophe) that is unjustified in what is simply a name. The Apostrophe Preservation Society wants to remove our ability to express such nuances in our naming.

So, Birmingham signposts. The only 'correct' solution is to follow common law - if most locals spell or punctuate a name one way then that can be the only arbiter. What Birmingham has found is that there is too much disagreement, so it is imposing a blanket rule for consistencey - no apostrophe. And this as well reflects a trend in modern English to minimise punctuation - if it doesn't make things clearer - leave out the punctuation (this blog - by putting language examples in italics, not quotes, follows this trend). But the nature of names allows too much variation to please everybody.

In summary, the rule for a possessive apostrophe applies strongly to common nouns where they have a useful grammatical function, but not so strictly to names, where there is a degree of variation consistent with (i) the freedom to give any name we like to an individual object, (ii) the varying degrees of descriptiveness and possession we feel is appropriate for that name. Having this ability to negotiate a compromise between competing forces in the language enables the uninhibited user to refine effectively and largely unconsciously their expressive power of the language that the self appointed guardians of 'correctness' would have us do away with. It is the Apostrophe Protection Society that is 'dumbing down' with its over simplistic and over-applied surface rules; if we were to accept their rules we would actually be restricting the English language, and breaking deeper and stronger rules that they choose to ignore because they weren't told about them at school. But that won't happen because these stronger forces win out over imposed or artificial rules. When speakers break the apparent rules, an intelligent linguist humbly sees this as a pointer to a wider set of rules and principles that are far richer, that are consistent with other areas of language, and that have far more enduring and satisfying power to explain than the petty rules that blinker the amateur prescriptive grammarian.

Below Birmingham: 'chaos'