Restarting a linguistic discussion about what a word is.

The late Professor Trask wrote an informative paper which gives four different definitions for the word word (link). Let's give it a read and see if we can come up with our own definition.

Trask's four kinds of words:

• orthographic words :- the simplest, "An orthographic word is a written sequence which has a white space at each end but no white space in the middle."
  
• phonological words :- "A phonological word is a piece of speech which behaves as a unit of pronunciation according to criteria which vary from language to language."
  
• lexemes (or lexical items) :- "A lexical item (or lexeme) is an abstract unit of the lexicon (vocabulary) of a language, with a more or less readily identifiable meaning or function." This is related to the citation form (or lemma) which is what's listed in a dictionary.
  
• grammatical word-forms :- "A grammatical word-form (or GWF, or grammatical form) is one of the several forms that may be assumed by a lexical item for grammatical purposes."
  

The third definition is the one which comes closest to what I consider a word to be. Some problems include: (1) if a word is not in the dictionary is it a word? (2) What criteria are used for separate entries in a dictionary? (e.g., the OED has separate entries for noun and verb forms of the same word, while A-H and M-W do not) and (3) What about compound words and idiomatic phrases? The final concern is perhaps the slipperiest slope of all. Sure, phrases like kick the bucket are often included in unabridged dictionaries because their meaning is not evident from the meanings of their constituents. Yet, do simple compounds like football make it in because of the lack of a space or hyphen? A simple phrase such as the Queen of England, while non-problematically intelligible from the meanings of the separate words still kind of acts like a word. Cf. forming the possessive, the Queen of England's throne, but not in forming the plural, the Queens of England. Then there are "words" like nonetheless.

Z, I'm glad you started this thread because when this was mentioned in another thread, I was thinking it belonged here. I have to print out the paper and read it. Early on here there was another great source on what is a word, and I don't know if the paper you posted is it or not. But I do recall that it was excellent.

I have noticed OED's separate entries for noun and verb versions of a word, and I don't like it. It's hard for me to consider different definitions of 1 word to be separate words. Yet, I can see the point because each definition has a different message.

We've talked here about what to do with words that aren't in the dictionary. Are they words? An easy definition, of course, is that the OED is the gold standard of dictionaries, and therefore if the word isn't there, it's not a word.

quote:

Originally posted by Kalleh:

It's hard for me to consider different definitions of 1 word to be separate words.

And yet...

What of words that have two or more extremely different noun meanings? Favourite example: set which has dozens. Is set (a varient spelling of sett) meaning "a badger's home" the same word, or a different word to set meaning "complete collection" or to set meaning "position" and so on?
Come to that is it the same word, or a different word to sett?

And then there are word pairs that are written the same but pronounced differently which can have related meanings as in the verb and noun pair produce/produce or unrelated meanings as in minute (1/60 of an hour) and minute (very small).

I'm sure we've discussed all this at length because I remember quoting almost a whole page from Crystal on the subject. The bottom line is that while one person may come up with a definition that satisfies him of what a word is it's extremely unlikely that two people will be able to agree on the subject, which is why the number of words is essentially not definable.

quote:

the number of words is essentially not definable.

And, ultimately, unimportant.

I note with some humor that you used "definable", not "determinable".

[Adding two links from goofy here: Ask a Linguist and Jesse Sheidlower.]

Many English dictionaries have separate entries for homonymous words based on their etymology, e.g., the many words spelled mole. Spelling influences some decisions though. For example, to and too have separate entries in modern English, but in Old English they were spelled the same, tō, and seemingly the same word (a preposition and an adverb) at least with the same etymology (link). (Two, though pronounced the same, is from Old English twā.

A question we'll probably be trying to answer soon in the Saussure RG is: what is the difference between a word's meaning and its definition in a dictionary?

quote:

Originally posted by zmježd:


A question we'll probably be trying to answer soon in the Saussure RG is: what is the difference between a word's meaning and its definition in a dictionary?

And when you do I trust you will also be quoting Lewis Carroll: "the name of this song is called 'Haddock's Eyes'" and Alice's subsequent confusion over things and the names of things - as nice a demonstration of metalanguage as I've ever seen.

Duplicated from the jokes thread in response to Kalleh's post there.

quote:

Yet, they just have to be considered one word, don't they?

That's exactly the point. They don't have to be considered anything. That some people would say one word and some two for those distinct meanings of "tick" illustrates the impossibility of counting. Similarly with all those compounds, you say that you don't consider them separate words whereas I definitely do consider them separate words. This is why we keep on insisting that a) a sensible definition of "word|" is very difficult and may be impossible, b) any kind of count is probably impossible, c) counting the words is a meaningless activity anyway.

Probably the biggest variation in the counts would come from inflected forms

run;runs;running;ran - one word or four
and what about
runner;runners

If those are different forms of one word what about

be;am;are;is;was;were

if those are different then why aren't the ones for "run"?

I've read the links and the posts here and I wonder if there will ever be an accepted definition of a word. And perhaps it makes no difference, though I am wondering why we use the word (Ha!) so often, especially on this forum! So I checked out the definitions for word on the OED, and this one I thought was the best:

quote:

An element or unit of speech, language, etc.

12. Any of the sequences of one or more sounds or morphemes (intuitively recognized by native speakers as) constituting the basic units of meaningful speech used in forming a sentence or utterance in a language (and in most writing systems normally separated by spaces); a lexical unit other than a phrase or affix; an item of vocabulary, a vocable.
Sometimes used specifically to denote either an item of vocabulary in the standard form in which it is generally cited in a dictionary, etc. (e.g. the infinitive of a verb), or this form considered together with its grammatical inflections as expressing a common lexical meaning or range of meanings.

I especially liked that comment about "intuitively recognized by native speakers" because I think that answers some of those questions in goofy's link (by z) where they talk about French, Turkish, Chinese, Japanese, etc. words. Why should word mean the same in every language?

Sheidlower's comment about their being 84 million chemical substances also gives one pause. We will never know how many words there are, I am convinced of that. Just in my field, medical diagnoses, instruments, procedures, drugs, chemical compounds, etc., change frequently; that's just my field! But, I do think we could come to an agreed upon...general...definition. It wouldn't be specific, but we could come to something, couldn't we? Something similar to that OED definition?

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

But you see that "intuitively recognised by native speakers" is another Get-out-of-jail-free card (rather like my art definition). I accept it as entirely valid but it's intrinsically useless to anyone seeking an objective definition (perhaps with a view to counting the words) even within one language. It amounts to "it's a word if people say it's a word" and that too is quite right and proper because, as I said before, you probably won't find two people who actually agree for all examples that something is a word.

"intuitively recognized by native speakers"

If you can't come up with a definition of word that's valid amongst many languages, how can you compare the number of words between those languages? It's all a bit like saying you know the highest number. You say googolplex, and the other person responds googolplex plus one.

And after the problems of what a word is, of counting words in a language, you're left with how the number of words in a language is some metric of that language's betterness.

Sheidlower's comment about their being 84 million chemical substances also gives one pause.

84 million named chemical substances.

quote:

84 million named chemical substances.

Reminds me of the American who asked the Frenchman how many sexual positions he knew. The Frenchman said, "Seventy-one, Mssr."
The American whistled and said, "I only know one, the man on top."
And the Frenchman slapped his forehead and said, "Sacre bleu. Seventy-two."

There's always one more thing to think about.

I think I mentioned that the number of words isn't important to me. But in English, at least, I would think we'd be able to define (generally) what a word is. If it doesn't work for every, single situation, so what? However, if we can't come up with a definition for word, I really don't see how we can be a forum about words. Oh well. I guess I am going around in circles.

At any rate, I like that definition from the OED that I posted above. That's my definition.

I think this discussion gets at the heart of what makes linguistics a particularly difficult subject for the layman-- what has attracted us to setting up a forum and trying to understand it better. Any field of study where the focus is an element of human culture-- (and what could be more intrinsically human than speech?)-- is bound to be resistant to the probing of the scientific mind. The elements that surround us in the natural world seem to have harder edges-- to stay put longer, or at least long enough to nail down a few definitions...

Yeah, deciding what makes a word is very difficult in any language, but it gets harder in languages that don't operate in the same way as English does. Agglutinative languages (link) such as Bantu languages make it very difficult to decide what a word is, since what in English would be an entire sentence can be expressed in what must be written as a single chunk and is made up of non-free-standing morphemes, i.e. can't be broken up into smaller free-standing parts. For example, 'Ngiyahamba' = 'I go' or 'I am going'. But splitting it up into its constituents: 'ngi-' is the First person marker (I don't know the correct linguistic terms), not quite like a pronoun because it cannot stand on its own. '-ya-' marks the present tense of an intransitive verb and '-hamba' is the stem meaning 'go'. However, you cannot say any of those parts on their own. It can get more complicated, with phrases like 'Ngizokubona' = 'I will see you' where '-zo-' marks future tense (roughly the equivalent of 'will') and '-ku-' is a second-person object prefix. '-bona' is the verb stem meaning 'see'. So to get a truly general definition of a word we need to consider the structures of different languages.

As for English, I would consider some words that look and sound the same to be different words. Can't think of any good examples right now though...

I'd call the ngi- in ngiyahamba a prefix. Linguists distinguish between bound and free morphemes. isiZulu ngi- would be a bound morpheme because it cannot stand on its own; English I would be free as it can. I have a friend from Lesotho who speaks Sesotho (a Bantu language). He lives in France now and teaches English.

Did the old chestnut that "The Eskimos have XX words for snow" possibly arise because of the agglutinative nature of their languages?

Did the old chestnut that "The Eskimos have XX words for snow" possibly arise because of the agglutinative nature of their languages?

No, it pretty much had to do with the Inuit living in a pre-climate-warming Alaska and environs. The original amount (in one of the works of Franz Boas) was four words:

quote:

[A]ll Boas says [in his Handook of North American Indians (1911)], in the context of a low-key and slightly ill-explained discussion of independent versus derived terms for things in different languages, is that just as English uses separate roots for a variety of forms of water (liquid, lake, river, brook, rain, dew, wave, foam) that might be formed by derivational morphology from a single root meaning 'water' in some other language, so Eskimo uses the apparently distinct roots aput 'snow on the ground', gana 'falling snow', piqsirpoq 'drifting snow', qimuqsuq a snow drift'. Boas' point is simply that English expresses these notions by phrases involving the root snow, but things could have been otherwise, just as the words for lake, river, etc. could have been formed derivationally or periphrastically on the root water. [G K Pullum The Great Eskimo Vocabulary Hoax and Other Irreverent Essays on the Study of Language, pp.162f.]

It would be fair to point out here that Professor Pullum is not the person who researched and debunked the snow clones meme. That was Laura Martin, an anthropologist at Cleveland State University in the 1980s. He simply popularized it in his book.