Job Discussion Forums

[naturegirl321]

About how many words does the average English speaker know? I've been asked this lots of times last week.
I remember when I was younger, reading something like the average teenager 100odd years ago knew 40,000-50,000. BUt now, because of tv and dumbing down, or so the article said, the average teenager only knows 10,000.
Anyone know?

[Roger]

I think, quantifying the unquantifiable is a very amateurish pastime proper for our Chinese English learners.
First of all, you will need a rigorous and reasonable definition of words. The same spelling, but a different category of words - say, noun versus verb - makes it two words, but in the mind of a typical CHinese person it's one word. Example: Choice: Many CHinese say "I choice ice-cream", or "I chose, I make a choose"...

Secondly, we don't "know" but "feel" the words; they come to our mind rather spontaneously. A native speaker needs to "know" specialised vocabulary such as metalinguistic terms ("noun", "verb") or proper of a professional ("scalpel"), etc. These specialisations don't normally enter the "lists" of regular vocabulary that our students learn.
Many words are in fact passively mastered; you may recognise a word such as "horripilifying" when it's uttered in a meaningful context, but you may never be able to utter it yourself.
I guess we all have several tens of thousand lexical units in our active memory, and potentially several hundred thousand passive units.

[Will.]

This may be of heklp to you Naturegirl.

It is from a programme on the BBC World Service or Radio 4, so a recognised source gave the numbers. I can't for the life of me remember who.

At age 5 the child has a range of +/_ 500 words. This rises to 50,000+ at 20/25 years.
A university student 's word knowledge equates with 2/3 of the Oxford Concise Dictionary i.e 75,000 words.
Between 5 and 20 years we acquire about 10 words a day i.e about 3,000 a year, but not at a smooth rate.
Bearing in mind all the related and specific jargon of our studies and trades there is much to be said regarding the knowledge of words and the continued use of them in our daily lives. the more varied our lifestyles the more....
I also read that the average 'Sun' (a tabloid newspaper here in London) reader has a base range of 2,500-3,000 words used for an average week.
The reference here is, again, to 'knowing' a word and to using it correctly on a regular basis as part of your personal lexicon. As Roger mentioned many words are passively mastered, we are not all capable of defining them as Dickens would have had us do in the style of 'girl 20'.
Hope that helps you.

[guest of Japan]

I read somewhere long ago that the average native speaker knows between 25,000 and 30,000. Those with higher levels of education are between 40,000 and 50,000.

[nomadder]

Know? 10,000
Use? About 100.

[Glenski]

On the Adult Education forum, you can find this thread.
http://www.eslcafe.com/forums/teacher/viewtopic.php?t=1223

[naturegirl321]

Yep, the question does come from CHinese. Thanks for all the info. It'l be daunting for them

[Atlas]

Shakespeare used 30,000 words in his plays.

The average adult needs to know about 2000 to communicate most needs in a foreign language. One should know about 3000 in order to understand others.

It is estimated that the average adult native speaker knows 11,000-20,000 words, but you won't find much more than 3000-5000 on TV. 5000 and above gets into university-level language (think, GRE vocab tests) and technical words that are industry-specific.

One of the problems about word counts is that everyone counts them differentl. Are [change, changed, changing] different words, or the same? There is no standard. There are millions of words in the English language, and more are added every year, and others don't pass the test of time ("Metrosexual"? Are you kidding me?)

It is hypothesized that the reason people know so many more words than they use is to show off in mate attraction, intelligence being one of the 2 most important factors (and kindness being the other). But these are psychological theories, subject to empirical revision--not written in stone.

Your students probably know enough words to be fluent--but they are not fluent because they refuse to use language authentically, and instead just spout textbook examples in an attempt to placate the teacher. Also, it's impossible to take a passive role in learning and become fluent; it really takes an active effort, and willingness to take risks saying something embarrassing, or worse, actually engaging someone in a foreign language and having to think on your feet. That's work, and it's much easier to shut up and lean against that big language barrier and have a cigarette. Anyway, that's the China situation. I've had students claim to know 10,000 words, but could not ask me for a cup of water if their * was on fire.

Hey, just thought of Monday's lesson plan!

[kimo]

I took the original query to heart and decided to count just exactly how many words I know. My count ended at 342.

So the answer for me is I know exactly 342 words.

Hope that helps!

[khmerhit]

Are you not a native speaker, Kimo?? (OK, """""native speaker"""""".)

Just curious, or are you joking us?

khmer

[Dr.J]

There are various estimates given by various sources, and it's hard to verify a method, as even when you don't know a word you can guess it's meaning fairly correctly from the context.

In any case, it is in the tens of thousands. I heard that English has a larger number of words in use than many other languages, but this may have been a lie.

Also it's important to distinguish between words we can understand and words that we actively use. Usually the former is twice as much as the latter. So when it comes to being a fluent sopaker, it's better to learn how to manipulate a small number of words very skillfully rather than learning a huge amount of vocabulary. Though vocab helps, of course.

[randyj]

I. S. P. Nation in "Learning Vocabulary in Another Language" talks about word families instead of just words. A noun and its plural might form a family, for example, or a verb and all its conjugations. According to Nation, language users probably need 15,000 to 20,000 words to read with minimal interruption. On the other hand, someone with the 2,000 most frequently used words in English plus the 570 word families in the academic list would know about four words out of five in an academic text, not considering specialized vocabulary. See also http://www.edict.com.hk/textanalyser/ .

[arioch36]

Took the GRE's about five years ago. Scored in the 99th percentile (Allright, hold your snickers) Been teaching in China, am now down to the 95th percentile. Give me a few more years, And I will be speaking English like a native.


Question?

DO I still know as many words?

The concept of "knowing" words is flawed. If we see or hear a word even once, it is permantly etched into our brain. We never really "forget " a word.

In academic english the key is in being able to correctly deducing the meaning of an unfamiliar word. We can use Slat's random generator to produce the words.

How many spoken words can we understand is important. If I know the word when I read it, but can not understand it when I hear it, do I know the word?

And what about phrases and idioms? If I know two words of the phrase, but don't understand the phrase, do I know the words?

Enquiring minds want to know. Me, I am going to watch CCTV 9 (China's English TV ) so I can become reconvinced of my English superiority, and so I can refresh my knowledge of Chinese english, so I can understand my students. Spent the weekend with laowai.

[naturegirl321]

[Ludwig]

As a �professionally trained� linguist, (in the sense that I have formal, academic qualifications in both general and generative psycholinguistics - BA (Hons) Language Studies and Linguistics; MA Language Acquisition; Ph.D candidate Evolutionary Computational Linguistics) the question �How many words does the average person know� is, along with the other usual suspect of �How many languages are there in the world� a question that I am asked surprisingly frequently.

Such questions used to make me grind my teeth and sigh, now I politely try to explain why, in fact, it�s a question, contra any immediate intuitions, that is far from simple to answer. To me, as I�m sure it is to any other professionally-trained linguist, the question posed above, is, in fact, a typically lay non-question (and I use �lay� here in its technical, inoffensive sense). I�ll try to show why in (what I hope are) easy to understand terms and arguments.

If you ask someone � anyone, even a layman � whether or not they believe they know what the word 'word' refers to, they are sure to (rightly) answer in the affirmative. Surely everyone knows what a word is?

Well, to a linguist at least (and as I said, I�m one of them), and especially to those engaged in researching �how many words people know� (again, I�m one of them), the answer is not so obvious.

It's a question that has been an active area of research within many various fields for some time. In psycholinguistics, for example, the study of the lexicon has attracted a considerable amount of research and even boasts of its own topic-specific academic journals. One thing you must be cautious of, however, is professional, scientific research conducted to normal levels of academic quality and procedure, producing specialist, technical data usually in concert or conflict with a particular theoretical framework, being then simplified and paraphrased for the non-specialist (pitched at about a �popular science� level) and then again being further simplified to a tabloid level in an attempt to make it somehow accessible to the general public. (This is what I suspect has happened to some of stuff alluded to in some other responses.)

As speakers of a language such as English, we may very well easily and casually analyse spoken language into what we call �words�, as evinced by our ability to, say, segment streams of speech into words and, if called upon to do so, can transcribe spoken words with conventional spaces in between, thus showing their status as �words�. Words are the entries in dictionaries; they are (among other aspects of the language) �things we teach� in the classroom; they are perhaps the parts of our language we talk about most often. We even often argue over the �exact meaning� of words. To us, surely, the word is the elemental building block of our language.

However, it might surprise you to learn that, to a linguist at least, whatever the language being analysed, the word is a poorly understood notion � at least in the sense of explicit definitions � and is most likely not the �elemental building block� of any language (and it is far from clear what, if any, feature of any natural language could objectively be considered as such).

Imagine if I were to ask you to conduct some research, to ascertain �how many English words your students know�. Now, leaving aside the question of what cross section of your students you will examine (this is a statistical question and need not detain us), you must establish what you are to measure and how you are to go about the necessary measuring (this is basically what the poster named 'Roger' touched upon). That is, how can you count how many �words� one knows until you have defined the word �word�. And how will you conduct tests? Will you run through a dictionary with your subjects and ask them if they �know� the word? How will you establish if they really do or don�t �know� the word? By, perhaps, their being able or not being able to, perhaps, give a dictionary definition? And what about the words used in their reply? Will you ask them if they 'know' those words? If not, how do you 'know' they are not merely repeating sounds in parrot fashion? Does merely repeating a word parrot fashion entail your 'knowing' the word?. Or perhaps you will say that a person 'knows' a word by their being able to place the word in an appropriate structure. How will you go about defining and clarifying what you will accept as 'an appropriate structure'? How would you go about clarifying what genuinely �knowing� a word is? That is, how will you go about defining and explicating what �to know a word� entails? How would you decree whether or not someone 'knows' the word 'gold'? By asking them to point to something gold in colour? Or would you want a chemical equation for the substance gold? What if they are fooled by Fool's gold? Will you, for example, count 'fingerprint' as one word or two? What about 'finger mark'? Will 'textbooks' be one 'word' and 'text materials' two 'words'? Are 'seaplane' and 'seaside', for example, one word? They look nice enough on the printed page or screen, yet 'seaair' and 'seaurchin' look, to most it seems, downright terrible. Will they be one word or two? Will 'common sense', 'black sheep', 'top hat', 'high school', and 'hot seat' be considered �expressions� or �idioms� and not words, or both? If so, why? If not, why not?

A German discussing English may very well insist that 'youth hostel' and 'traffic accident' are single words, as indeed they are in his native German, 'Jugendherberge' and 'Verkehrsunfall', respectively. If you were to examine a Japanese informant you will most likely encounter similar difficulties. The word closest to the meaning of �word� ('kotoba') in Japanese can, in fact, be used to refer to ANY meaningful expression: a monosyllable, a phrase, a saying, a statement, or even the language as a whole. To those of you with some knowledge of Greek, you will know that the verse that is the opening of the Gospel of St. John, viz., �In the beginning was the Word,�� sounds somewhat strange and mysterious because �Word� imperfectly translates the term 'logos'. But to return to the thought experiment research I have asked you to do, you must establish (a) what you are to measure, that is, your definition of a �word�; (b) how you are to collect data, that is, your testing procedure; and (c) how you are to go about defining a criteria for what it means in this experiment to �know� a word.

These points are obviously not trivial. Of the three, (c) is the most theoretically challenging, and would take us well beyond this question if that were to be discussed to any decent degree at all (though as we have already seen and shall see again it cannot be totally avoided). Let�s, then, focus on (a) and (b). Let us also stick with English.

One methodology is as follows (and I used this for a module in Second Language Learning that was part of my MA ). Entries are taken from a large unabridged dictionary, including, of course, only those entries, the meanings of which can not be deduced through derivational morphology or analogy. This is vital since even if you had never heard, say, �restart� you could guess what it means and so it would no longer be a test of �how many words� the person �knows� (in the sense of how many words have been learned which, actually is the question I suspect you actually mean). Since, of course, it would take too long to test a bank of people on hundreds of thousands of words, a random sample is instead taken. The proportion of that sample that the informant in the test knows is recruited to generate an estimate of their overall vocabulary size (under the assumption that the dictionary used is a reasonable estimate of the language as a whole. So, if you employ a dictionary with, say, 500,000 entries, and test people on, say, a 500-word sample, you determine the number of English words they know by taking the number of their �accepted�, �correct� answers (again, you will have to adopt a methodology on this point as to what, exactly, constitutes an appropriate answer) from the sample and multiplying by 1,000. The 'meaning' side of the word coin is usually approached through a multiple-choice question with four or five alternatives. (This of course introduces a pure chance factor that must of course be controlled for, but again this is simple enough to do statistically.)

Through such random sampling techniques (and the above methodology appears to be due to Miller, G. A. (1996) 'The science of words'. New York: Freeman) it has been estimated (Nagy, W. E. & Herman, P. A. (1987) 'Breadth and depth of vocabulary knowledge: Implications for acquisition and instruction'. In McKeown, M. G. & Curtis, M. E. (Eds.) 'The nature of vocabulary acquisition'. New Jersey: Erlbaum) that American high school graduates, for example, have a lexical store of about 45,000 unique listemes.

However, as noted by Paul Bloom in his 'How children learn the meanings of words' (Massachusetts: MIT Press, 2000, p. 6), this does not include proper nouns for individuals and places, idiomatic expressions, and undecomposible compounds. Once such words ARE taken into account, this estimate increases to between around 60,000 and 80,000. This figure is analysed and deemed worthy of being a profitable working estimate by, among others, Aitchison in her 'Words in the mind: An introduction to the mental lexicon' (1994, 2nd ed. Oxford: Blackwell), and Pinker in his 'The language instinct' (1995, London: Penguin). For polyglots boasting of equal lexical competence in all their languages, of course, such figures need to be multiplied by the number of languages spoken.

Since kids begin to produce words at around the age of 12 months, some quite elementary maths shows that, if we stick to the conservative estimate of 60,000, this lexical store equates to the learning of around 10 new words a day from 12 months up till the end of High School. This, in turn, equates to around 1 word every 90 minutes (spent awake). This is not to be sniffed at. How many of you learners of Chinese learn ten words a day, day-in, day-out? In the popular science introduction to Chomskyan generative linguistics referred to above, 'The Language Instinct', Pinker refers to young kids as �lexical vacuum cleaners�. The description is apt. He asks us to remember that words are Saussurian signs and asks us to imagine having to try to learn 10 other arbitrary pairings such as batting averages or treaty dates or 'phone numbers every day.

So what is a �word�? Well, what IS it? One notion popular in linguistics is that the word is a syntactic atom, something that can be a member of a word class, such as, for example, Noun or Verb, and that can be the product of morphological rules. This is certainly close to our intuitive notion of a word: a sound or a sign that, if written down, corresponds to a string of letters that has spaces or punctuation marks on either side (though even here you have some not inconsiderable problems: Sanskrit, for example, was - and still is - written as one long string).

Under THIS definition, the sentence, �John stayed in the poker game until he got cleaned out� has 11 words. (This assessment is confirmed by the word count tool on my PC when I copy and paste this sentence to Word.)

However, this notion of word is totally unsuitable for linguists interested in �how many words people know� (with �know� in the sense of �had to learn�. As a child you did not have to learn the word �stayed� and adult learners of English as a second language do not have to learn �stayed� either. What you did have to do and what they still (sometimes) have to do is to learn the verb �stay� and the morphological rule that adds (what is orthographically represented as) �ed to transform base form verbs into the (in this case) simple past tense. (Also, think about the adjective �staid�. Would that be yet another word all because it just happens, through nothing more than historical accident, to be spelt differently?)

The above example of �John stayed in the poker game until he got cleaned out� also harbours another problem; namely, that of idioms. To understand this sentence it is clearly not enough to know the verb �clean� and the preposition �out�, but, rather, have to know the meaning of the idiom �clean out�, which means roughly something akin to being totally deprived of something, usually money. Consider also the seemingly innocent noun of poker. From a �know� perspective, this string of letters is obviously (at least) two wholly separate words: a card game (one word or two?) and a fireplace (one word or two?) and each must be known separately. It is clear then that to make any headway towards a definition of a �word� there must be some reference to meaning. When you see the results of empirical research that have been watered down to a popular science level, or perhaps even to a tabloid level, and see that �linguists x,y, & z say that the �average speaker� �knows� x words�, or that �Writer x used x number of words�, what has usually happened is that the linguists have actually referred to lemmas; that is, a listing in a dictionary. This has the clear advantage that �stay�, �stays�, �stayed�, and �staying� are counted as a single word, but the disadvantage that �poker� counts as but one word, despite its ambiguity when confronted in isolation.

Anyway, so in short, it's a (heavily) qualified figure of between 60,000 and 80,000. (Though I would argue that, since the majority of the world's population is estimated to be (at least) bilingual, the 'average person' may, in fact, 'know' between (at least) 120,000 and 160,000.)