|
 |
How many English words do you know? | ||
Share with:
| ||
 Language Learning Forum : Specific Languages |  | |
| mrwarper Diglot Winner TAC 2012 Senior Member Spain forum_posts.asp?TID=Registered users can see my Skype Name Joined 5035 days ago 1493 posts - 2500 votes  Speaks: Spanish*, EnglishC2 Studies: German, Russian, Japanese | Message 17 of 21 30 January 2014 at 10:15pm | IP Logged | ||
I haven't taken the test, but from what I've read, I think that people get too carried away with their mathematical mumbo-jumbo before getting the basics right. If we go to some levels of refinement, the problems with defining what is or isn't a word are not negligible either, and they may add more noise to measurements. Or so I think. Let's start by saying 'words are in the dictionary, and non-words are not'. The 'problem' with words is people can guess and say something is a word even if they're not sure. Non-words are tested to compensate lucky guesses in finer measurements. But there are different kinds of non-words in real life, and people will respond to that in different ways. Random strings are likely to produce zero false positives in non-words, sure. Random syllables that appear in English, combined in ways that don't appear in any dictionary? Maybe, maybe not. But presenting 'non-words' that are actually understandable is opening a can of worms. In real life, people use words meaning something different from what the dictionary says all the time, and they get away with it as long as they are understood and they don't bump into a 'dictionary nazi'. For the same reason, people make up words with varying degrees of success, which hints at how words and non-words are not that clearly disjunct sets in real life. Is this taken into account in the test? Are rare and obsolete words to be discarded as non-words? How would the test taker know? And if they did, how would they know which rare words should be marked as non-words? Most importantly, are test takers supposed to react to non-words in the test in the same way they would in real life? After all, in real life, test takers would only 'discard' what isn't understandable or whatever grates their ears. 2 persons have voted this message useful | |||
| josepablo Tetraglot Senior Member Portugal Joined 3799 days ago 123 posts - 141 votes Speaks: German, Spanish, Dutch, Portuguese Studies: Russian, Mandarin, Turkish | Message 18 of 21 31 January 2014 at 12:43am | IP Logged | ||
|
First try: I scored 90% of the existing words, 0% of nonwords.
Second try: I said yes to 91% of the existing words. and yes to 3% of the nonwords. "This gives you a corrected score of 91% - 3% = 88%." I said yes to the nonword douseness, which sounded quite convincing to me. The 6 words I did not know: agenesis, agnatic, angostura, disodium AND two I did know: rocketry and arterially. Don't know why I pressed F, must have been getting bored. 1 person has voted this message useful | |||
| dampingwire Bilingual Triglot Senior Member United Kingdom Joined 4474 days ago 1185 posts - 1513 votes Speaks: English*, Italian*, French Studies: Japanese | Message 19 of 21 31 January 2014 at 1:24am | IP Logged | ||
The first three are specialised terms. disodium isn't really a word, just something that crops up in chemical names ... some of those part-names I'd count and some not (dioxide I'd count, but I don't remember ever seeing disodium except in a chemistry lesson). I just had a go and got 87/0. The ones I failed to identify as English words were legal terms, plants a drug and provenience (apparently a US variation of provenance). I noticed a fair number of chemical terms in the list that I did know. If I'd swung towards the arts at school that would have significantly hampered me on this test (at least for the words that came up for me this time). I took the longest over "seeped" (3.482s) and the quickest was "son" (0.662s). There's clearly some kind of competition waiting to spring to life here :-) 1 person has voted this message useful | |||
 Iversen Super Polyglot Moderator Denmark berejst.dk Joined 6512 days ago 9078 posts - 16473 votes Speaks: Danish*, French, English, German, Italian, Spanish, Portuguese, Dutch, Swedish, Esperanto, Romanian, Catalan Studies: Afrikaans, Greek, Norwegian, Russian, Serbian, Icelandic, Latin, Irish, Lowland Scots, Indonesian, Polish, Croatian Personal Language Map | Message 20 of 21 31 January 2014 at 2:44pm | IP Logged | ||
|
I have followed the link to the d'-test, but on the machine where I tried it out the d'value always turned out to be 0.000 - maybe the test is java version dependent? And the mathematical intricacies of information theory are beyond me. But I have grasped the general idea - that you shouldn't be able to optimize your score by either avoiding guesswork or guessing wildly. And some of the remarks in the research on the page which patrickwilken referred to indicate that you check the guessing-propensity of a testee on one side by comparing his/her scores to scores based on totally random guesses and on other other on a 100% perfect knowledge of the correct answers (correct answer = a word on the list compiled by the researchers, whether it is in practical use or not).
The central problem is that the notion of 'existing' isn't totally black and white. Some of the fictive words in the ugent test look like simple variations on existing words, and if you used them with people who knew the relevant existing words they would understand your neologisms. Actually I often find words in texts which aren't in my dictionaries, and many of these are probably invented by the author. As I have written earlier I operate with a 'guessable' group in my latest word counts, and my goal should then be not to include non-existant words in the "exists" group - but it would be a bad sign to leave the "guessable" group empty. My own vocabulary estimates (last updated in Nov. 2013) are based on a selected pages in standard dictionaries, which means that there isn't a testgroup with constructed words which could be used to test the validity of the numbers. But now that I have a "guessable" category it turns out that it much smaller than the "known" and "unknown" categories - in other words: I mostly do know whether I know a word or not. And I have just checked the numbers: my own tests also however around 50% known words in Dutch, but in English my estimates from 2012-13 have an average around 69% (spanning the interval 64 to 78%), and I count 8% guessables (insofar such a word exists). But from four earlier sessions I got an average of 73%, which however covered a span from just 31% to 92% - so it seems that having a special box for "guessables" has stabilized my scores. And funnily enough my own estimates in English lie significantly below those I got with the online test. Edited by Iversen on 31 January 2014 at 2:51pm 1 person has voted this message useful | |||
| Hungringo Triglot Senior Member United Kingdom Joined 3797 days ago 168 posts - 329 votes Speaks: Hungarian*, English, Spanish Studies: French | Message 21 of 21 31 January 2014 at 2:47pm | IP Logged | ||
|
You said yes to 81% of the existing words.
You said yes to 0% of the nonwords. This gives you a corrected score of 81% - 0% = 81%. You are at the top level! In theory this means that my English vocabulary is close to 50 000 words. Edited by Hungringo on 31 January 2014 at 2:57pm 1 person has voted this message useful |
If you wish to post a reply to this topic you must first login. If you are not already registered you must first register
|  | 
Printable version
|
| Copyright 2024 FX Micheloud - All rights reserved No part of this website may be copied by any means without my written authorization. |
