I was all set to pounce on this nonsensical claim when I read the article and discovered it pounced rather more effectively than I would have.
David Crystal (who I had the pleasure of hearing speak at a lecture last week) in his book The English Language puts the figure at anywhere between half a million and "more than two million" depending on how you do the counting. There's quite a long and entertaining chapter on why counting the words in English isn't possible.

This makes me think of the study of how parents interact with children.

Meaningful Differences talks about the results of that study. They've found that the amount of speaking you do with a child (especially those conversations you have that go beyond mere custodial interactions) have a profound impact on their cognitive ability, particularly language skills (including reading).

Well worth the read if you are interested in such things.

quote:

in his book The English Language puts the figure at anywhere between half a million and "more than two million" depending on how you do the counting. There's quite a long and entertaining chapter on why counting the words in English isn't possible.

Ahhh, but you see, this all gets back to the question I've asked here since this board started: What makes a word a word? That the crux of this question. Is it any dictionary word? If so, epicaricacy would count as a word. Or is it only OED words? If so, then there are a lot of slang and newer words missed. Of maybe they aren't formally considered words.

You see, we'll never know, I think.

I have heard claims of up to 5 million words - but that includes scientific and specialist words. Out of interest I looked at OEDILF's list of undefined words and there are still 51 words beginning with "bi" that haven't been defined and the list's first 9 entires are:

bianchi
bidon
bilateral diagram
bimodality
binder bolt
biocompatible
bioenergetic
bioethical
biogenetic

and so on. I suspect that some of this random sample wouldn't be included in most counts - certainly not all the words appear in my COED.

quote:

Originally posted by Kalleh:

quote:

in his book The English Language puts the figure at anywhere between half a million and "more than two million" depending on how you do the counting. There's quite a long and entertaining chapter on why counting the words in English isn't possible.

Ahhh, but you see, this all gets back to the question I've asked here since this board started: What makes a word a word? That the crux of this question. Is it any dictionary word? If so, epicaricacy would count as a word. Or is it only OED words? If so, then there are a lot of slang and newer words missed. Of maybe they aren't formally considered words.

You see, we'll never know, I think.

And as I'm sure has been noted. It's a "how long is a piece of string" question, intrinsically meaningless.

As both that article and David Crystal point out it can't be answered in any meaningful way.

Is flowerpot a word or should it not count because some people write it as flower pot?

If "run" is a word then are "running","runs", "ran and "runner" words? If some of them why not all of them? If some of them aren't then Why aren't they?

Is "high" the same word in the phrases "high time", "high hopes", "let's get high", "high society"?

What about "badger" the animal and "badger" the verb?

I suppose it word be possible to construct a question along the lines of "How many distinct combinations of letters form uninflected forms of words in the ENglish language regardless of how many meanings each of those combinations of letters have." But what would be the point? What would the answer actually show?

quote:

Originally posted by Kalleh:
Is it any dictionary word? If so, epicaricacy would count as a word.

You've seen in in a predecessor of the OED, so you know it's a word! Only it was spelt with Ks, not Cs.

Ahhh, but you see, this all gets back to the question I've asked here since this board started: What makes a word a word? That the crux of this question. Is it any dictionary word? If so, epicaricacy would count as a word. Or is it only OED words? If so, then there are a lot of slang and newer words missed. Of maybe they aren't formally considered words.

OK, it's a word. It's just not spelled correctly in Bailey's as noted by Bob. Nor is it spelled correctly according to tht only quotations in literature that have been discovered: in Aristotle's Nicomachean Ethics and Burton's Anatomy of Melancholy, both of which places it is spelt in Greek letters. But it's a word, just like humungous and gianormous are. Also, since we're counting Greek words, let's just take over the entire Greek vocabulary. And Latin, too. Brung is also word. Bring, brang, brung. And, schadenfreude is a word. A more commonly known word than the e-word (which cannot be used in writing or conversation without resorting to footnotes or an attached PDF of Bailey's Dictionary). My problems with the e-word are esthetic, historical, and personal. In fact, I think Mrs Byrne probably just mistyped it in her book.

[Fixed one citation.]

This message has been edited. Last edited by: zmježd,

Yes, you can't be exact – but much of this can be simplified down if we're looking for a rough count and not an exact one. Many of the points are either relatively small change, or can be answered by standard dictionary conventions.

• Since you'll sometimes see flowerpot written as two separate words, is the combined form a distinct word?
  Response: Most compounds (e.g., basketball) are spelled only as a compound. There aren't enough 'sometimes' items like flowerpot to matter, when we're talking about hundreds of thousands of words.
  
• Do you separately count obvious grammatical variants of the "same" word: two for most nouns (cat and cats); four for most verbs (I count; he counts; I counted; I am counting), and the -ly adverb from each adjective?
  Response: Follow the conventions of whether a dictionary lists the forms under a separate lemma.
  
• What about non-grammatical multiplication of forms (run and runner), or compounds with a distinct meaning (high times), or multiple meanings of a single spelling (sometimes as the same part of speech, sometimes not, as badger, noun and verb)?
  Response: Use dictionary conventions.

With those items taken care of, what other factors could significantly affect the count?

This message has been edited. Last edited by: shufitz,

quote:

[diversion re 'epicaricacy'] the only quotations in literature that have been discovered: in Plato's Nichomean Ethics and Burton's Anatomy of Melancholy, both of which places it is spelt in Greek letters.

Don't forget the Trench quote that OED gives. (P.S. It was this board that uncovered the Burton cite and, I believe, the cites to Aristotle's ethics.]

Well, I read about the Aristotle and Burton citations on the web and blogged about them in January and February 2004. (I never found the page in Burton, but I did find the Aristotle page.) You posted about the Burton citation on this board in November 2005.

quote:

And as I'm sure has been noted. It's a "how long is a piece of string" question, intrinsically meaningless.

Well, I disagree. If we were interested in doing this (and I don't really see the value in it), you'd just have to define your terms. That is:

1) A word is a word; "badger," no matter what meaning you use, is one word.
2) The OED is the gold standard for words. If a word appears in it, it counts. If not, it doesn't.
3) etc., until all questions are dealt with.

Then you could have your list of words, and the only thing left would be to count them!

But the problem is that if you take a word - to use an old favourite - like "set" it has pages and pages of meanings in the OED listed under more than a dozen main entries with dozens of sub-entries for each.

OK we'll take "set" as one word irregardless (and there's another kettle of fish ) of whether it has multiple meanings.

Now, do we include "sets" or "setting"? I don't have the dictionary to hand but I believe "setting" has its own entry while "sets" doesn't.

If we include the present participle form "setting" (because it does have an entry) do we include all the others for every verb whether or not they have a separate entry (some do, some don't).

Take a regular verb. Play? Yes. Plays, playing, played? Yes if they're "in", no if they're "out"? Yes anyway? Why?

As I said in my OP you could construct an enquiry along the lines you suggest or you could simply count the headwords (or maybe distinct headwords) in the OED but what would it actually prove? I can state with mathematical certainty that the number of possible letter combinations for words of eight letters or less is 26^8 + 26^7 + 26^6 + 26^5 + 26^4 + 26^3 + 26^2 + 26 which is, unless my maths has gone awry, 217,180,147,158 and that of those the vast majority don't form pronouncable sequences but so what? One question, is, I suggest, as meaningless as the other.

While I mostly agree with the "How long is a piece of string?" point, I think there's a better way to put it(and this question is not original with me): How long is the coastline of England/Scotland/Wales?

The answer, of course, depends on the length of your yardstick, with the upper bound (when your yardstick is down in the nanometer range) being veeery high. This comes from fractal theory, but it applies to the "how many words in English" -- especially when it comes to trying to nail down the number precisely, you find that your target won't stand still. In the case of the coastline, at a certain attempted level of precision, you'd come to realize that the tides are a factor, for instance.

But that doesn't mean we can't come up with a reasonable answer in both cases; you just have to live with the fact that it's gonna be a ballpark estimate; and it still lets you compare word counts and coastline lengths with other languages and countries.

David

The lemmatization problem increases in a language like Finnish where a word can have thousands of forms, most of which are regular, and all but disappears in a language like Mandarin Chinese, which has almost no morphology. Most dictionaries include cross references—not really sure you could call them a full-blown entry of irregular forms—"dying, see die verb".

Most dictionaries have an editorial process by which one could determine how many words a certain edition of the dictionary contains, but that would never equal the number of words in the language.

Froeschlein notes:

quote:

"how many words in English" -- you find that your target won't stand still. But that doesn't mean we can't come up with a reasonable answer; you just have to live with the fact that it's gonna be a ballpark estimate;

Precisely. There are dozens of factors that prevent an exact count but don't loom so large as to prevent one from making a reasonable estimate. To make that estimate, we need deal only with the "big-ticket" factors.

As noted above, we can use dictionary practices to deal with whether two items represent the 'same' word, or two distinct words. (I assume this is what zmj means by "lemmatization".)

What other factors could have a significant impact on the count? So far, I don't see that we've named any.

I linked the word lemmatization to its Wikipedia article which defines it. Ditto with morphology.

Other significant impacts are:

1. Dialect, jargon, and taboo words. I've mentioned in a thread how appendicitis was not included in the OED at first because it was against policy to include rare medical terms. This changed when the Prince of Wales had it, and it got big coverage in the media of the time. The difference in word count between the 2nd and 3rd unabridged Webster's had a lot to do with taboo words and non-standard ones making it in. (And there was quite a ruckus raised when this happened.) The OED included some dialect words, but not all, because there were other dictionaries being comiled at the time that addressed those regional and dialectal words.

2. Compounds whose meaning can be determined from their constituent parts: i.e., blackbird is not just any kind of ablackbird, but a particular species, baseball has a meaning that cannot be determined from base and ball. The OED treats many idiomatic phrases that use the lemma under the entry, and not in separate entries.

3. Some dictionaries have separate entries for words which are homonyms but have different etymologies. Some don't.

4. I've seen a good example of multiple competing terms for technologically innovative products, where in the end only one will win out. What I call a memory stick, others call a flash drive or a jump drive (probably a TM term).

5. It's not a trivial process, but it is a well known one. Why for example is there only one entry be for the copula? Forms are am, are, is, was, were, being, and been. They all look like words to me, but they don't get counted. And are is a homonym: 1st person singular in questions (aren't I), second person singular and plural (your are), first and third persons plural (we are, they are). Five words or one?

6. And, I mentioned above all the words that people use on a daily basis that aren't in the dictionary. I use (jocularly) and have heard others use gianormous for 'really huge'. How about a friend who commonly says clumbersome. Is it a word? I know what he means, he knows, others? Many families have private vocabularies. Should they be counted?

[Addendum]

7. Many words have dropped out of usage that were not recorded. Should they, too, not be counted? Impossible, I know.

8. A language's vocabulary is not fixed. It can be increased or decreased at any time.

9. Some dictionaries include proper nouns, and some do not.

More factors...

Which English do we mean? Are we talking about British English, in which case the word "sidewalk" isn't a word but "pavement" is with an entirely different meaning to the US "pavement"?

Are we talking about US English, Australian English,Canadian English? What about Indian English - which includes words like ayah (nurse), lakh (a hundred thousand) and crore (ten million) - all within the language that they consider to be English. (And there are a significant number of them - certainly thousands.)

That's before we start on spelling. Are "colour" and "color" different words because they have different spellings are the same word because they have the same meaning?

What about creoles and pidgins of English? Should we count their, often very colourful, vocabularies as part of English?

What about slang and dialect words? In or out or only in if the OED decides they are in?

Can you really say that "bostin'" isn't a word but words which noone has used for several hundred years but happen to be in the OED are?

If we are going to count the OED as the font of all wisdom where does that leave MW. If a word is in MW but not OED is it not really a word at all?

I stand by my assertion that it's both impossible and meaningless to try to answer the question of how many words there are inEnglish.

To add to Bob's list: the OED decided to include Middle English words, but not Old English. Why? Old English was no less English than Middle English. (This only counts for Old English words that fell from use in English (both Middle and Modern); if a citation exists for an Old English form of a Modern English word, it is given, but it doesn't get counted as a different word though its form and meaning may have changed.)

[Adding on to my previous list.]

10. What does it mean that language X has 100K words and language Y has 25K words? Is X more expressive than Y? What about languages where the very word word is even more problematic than English. Many Native American (and First Nations) languages don't really have words that are separate from clauses or sentences. Roughly, this means that you can't really say rain or dog, but must say something like It is raining or the dog exists.

As a lawyer, I appreciate hairsplitting and quibbles as much as the next guy. But we have some hairsplitting and quibbles that are irrelevant to the general issue.

We want a rough count of words, and we're somewhere in the neighborhood of a million. In that neighborhood, any issue that could move the count by a few thousand, or even a few tens of thousands, is small change, a quibble we need not care about for our rough count. So can we drop those of the above points which, however interesting, wouldn't account for (say) 5% of the count?

And as to regionalisms: There are words peculiar to (say) Newfoundland, as we've seen, that aren't part of English in the broader world. Can we accept that these are indubitably "Newfoundland English", but aren't part of "English as a general idea"? There are over a billion speakers of English. I'd exclude, as non-general, anything confined to a national or regional group that accounts for less than 5% of the speakers.

Now with just those two points in mind, how many of the above "difficulties" remain unaddressed? If Bob and zmj will list them, and we'll see if we can address them too.

We want a rough count of words, and we're somewhere in the neighborhood of a million.

If you want to say there are 1 million words in the English language that's OK with me. I, like Sheidlower, don't think the statement has much meaning, but then I don't think much of sports stats either. It's like asking somebody what's the highest number. No matter what you say, I can always add one to it. There is no highest number, and there is no exact number of English words. Plus or minus 5% of meaningless is still nothing much. I see counting the vocabulary of English, or any language, to be like painting the Golden Gate Bridge. By the time you're finished, it's time to start again, but that's not to say we shouldn't paint it. So, paint on. Because, nothing that I, Bob, Jesse, or the man in the moon might say will disuade you from your task. Post your findings when you get them.

Would you not agree that English has (at any instant) a finite number of words? For example, surely you'd agree that it's less than a million million millions!

But that pulls the rug out from under your position. If the number of words in the English language, it is emphatically not "like asking someone what's the highest number".

The number is of course impossible to know precisely, and it may be difficult to estimate. It certainly grows over time (which merely affects the estimation process). Of course the exact number (unknowable) and the estimate depend on what we decide to include or exclude, but rules can be prepared for those determinations, and the count made accordingly. (The rules will not be 100% precise, but there is no basis for presuming that they cannot be precise enough to produce a decent estimated count.)

zmj, simple question: which of the factors you and Bob note would affect the count by more than 50,000? (Assume we've decided, as above, the question of forms of "English" to include.)

quote:

I, like Sheidlower, don't think the statement [of how many words there are in English] has much meaning

Sheidlower doesn't say any such thing. He simply says that before coming up with a number or estimate, you must "agree on the basics" and "define criteria" of what shall count.

No problem there. By analogy: to determine how many men over 30 live in the US, you must decide how to count 1. hermaphrodites; 2. persons born on Feb. 29; and 3. persons having multiple abodes or no fixed place of residence. Even with that done, the count will never be precise. But none of this makes the question "how many men" a meaningless question.

which of the factors you and Bob note would affect the count by more than 50,000

1. Say there are 500K words in English recorded in the standard dictionaries.

2. Say that 100 K of these words are verbs.

3. Not being sure whether to count just count, but also all its forms (e.g., counting, counts, counted) would affect the number by at least fourfold. So, are there 500 K words or 900K words? Other verbs may have more forms. Are these forms not words? Why aren't we counting them? Why should we?

Deciding whether or not to include proper nouns (they're words as far as I'm concerned) could change the number by factors of ten. (This is a wild guess, but not outlandish.)

Sheidlower doesn't say any such thing. He simply says that before coming up with a number or estimate, you must "agree on the basics" and "define criteria" of what shall count.

quote:

It's probably possible to devise criteria that would allow us to conclude that there are about a million words in English. (The dictionary publisher Merriam-Webster goes for "roughly 1 million words" in its discussion of this particular question, although elsewhere, they suggest that the figure could be many millions.) But there's no possible way to count the actual number of words in the language, and the idea of having a running counter, as is found on GLM's home page, is absurd. So, why have journalists fallen for the claim? I think it's the pseudo-scientific nature of GLM's "methodology": The company claims to use an "algorithm" called the "Predictive Quantities Indicator," so its figures must be right.

So, he says there's no possible way to count the actual number of words in the language, Good enough for me. I can edit my original rhetorical statement. It's an impossible task.

I had said, "Well, I disagree. If we were interested in doing this (and I don't really see the value in it), you'd just have to define your terms." In other words I don't see the value in it; therefore, I'm with Bob and Zmj on this. The only difference is that I believe you could do this, as long as you had a list of what to include and what not to include. Frankly, I don't even think it would be that hard.

quote:

OK we'll take "set" as one word irregardless

Bob, with all our discussions of "irregardless," how could you?!

P.S. As I had said earlier, "set" would be one word; never mind all the definitions. Then, the decision would be if "sets" and "setting" were separate words.

zmj, maybe our disagreement is chimerical.

Shiedlower says a precise count is impossible¹, but that a rough estimate is possible provided one first agrees on counting criteria². I don't disagree, and if I understand correctly neither do you. You were merely emphasizing the first part, and I the second. And of course, Shiedlower is quite properly demolishing the spurious precision of the supposed 988,968 number.

---

¹ "there's no possible way to count the actual number of words."
² "It's probably possible to devise criteria that would allow us to conclude that there are about a million words."

quote:

I don't disagree,

A double negative; how very British :-)

quote:

Originally posted by Kalleh:
Bob, with all our discussions of "irregardless," how could you?!

I wanted to see if I could manage to say the word without taking my tongue out of my cheek.

quote:

wanted to see if I could manage to say the word without taking my tongue out of my cheek.

How did you fare?

Irregardless is an interesting word. The other day, I paused in wonderment, when writing the word re-release, as in "I'll re-release it soon", in an email. Release does not mean to lease again, in this sense (as in perhaps, "I may release the Beemer, but right now I'm not sure". I've always enjoyed the difference in signage in the States and the UK: for lease vs to let.

quote:

Irregardless is an interesting word.

How so? Isn't it just a mistake that became common and therefore got published in the dictionaries? Since the "ir" and "less" are both negative, it's not logical. I know, words aren't logical...but still!

I do see that Quinion supports the word and thinks that Americans don't give it its due. He makes a good point, really, saying that it has a fine flow about it and that it has a "stonger negative feel." I suppose. He also cites precedents for double negatives. He could persuade me to change my mind.

While he seems to indicate that is those Americans who berate the word, I note that Richard doesn't like it. Perhaps, in reality, Richard is an American.

Isn't it just a mistake that became common and therefore got published in the dictionaries?

It's a word. This is the new me. People use it, and people know what it means. Dictionaries don't make words. They simply record as many as they can or want to. Kalleh: you can't have the e-word and refuse irregardless its wordhood. I think that a word has arrived when a prescriptivist starts ranting and railing against it.

quote:

While he seems to indicate that is those Americans who berate the word, I note that Richard doesn't like it. Perhaps, in reality, Richard is an American.

The word is not common in UK English. I don't recall ever having seen it except in US writing. That's not to say that it won't be adopted here but I see no real need for it myself.

I see that Jennifer Saunders http://www.write101.com has today written her article on the topic of synonyms, whcih she claims are generally a good thing - though I don't myself feel that regardless and irregardless are synonyms in the sense that she discusses them. Her point is that we have many synonyms that derive from different root languages, giving as examples, among many others:

anger wrath
raise rear
ill sick

But my feeling is that these add colour while "irregardless" simply adds error.

We have discussed this many times before. I found 52 matches when I searched the board for it.

Yes, irregardless is a word, like it or not. The AHD notes it's a nonstandard word and that many people don't like it. The OED Online defines it as "In non-standard or humorous use: regardless," and says it is probably a blend of irrespective and regardless. The first citation is from 1912: "WENTWORTH Amer. Dial. Dict." The second citation may be more appropriate here: "1923 Lit. Digest 17 Feb. 76 Is there such a word as irregardless in the English language?"

The Online Etymology Dictionary says the "erroneous word" was "perhaps inspired by the double negative used as an emphatic."

Tinman

This message has been edited. Last edited by: tinman,

quote:

Kalleh: you can't have the e-word and refuse irregardless its wordhood.

Well, Zmj, did you see how I changed my mind a bit during that post? As I had said, World Wide Words was very persuasive. And, Richard, Quinion is British and discusses the word.

Tinman, yes, we have discussed it before here, along with other subjects (we have cited the word "epicaricacy" 225 times here). Because people come and go on forums, that will always happen.