This page is no longer active. It is being kept for historical interest.
	No discussion is needed to revive this page; simply remove the {{inactive}} tag and bring it up to date.

This is a page of past discussions on Wikisaurus, starting in 2002 and 2003, getting more traffic in 2004, with most discussion ended by the end of 2006. This is not a policy page. For further discussions about Wikisaurus, see Wiktionary:Wikisaurus#Discussion.

See ] for a growing list of Wikisaurus entries.

• Have a look around at the entries, and the way they are referred to from other entries.
• Try adding your own entries in existing WikiSaurus entries
• Try adding your own WikiSaurus pages.
• Let me know what you think of it--Richardb 15:04, 6 Dec 2004 (UTC)

• It is only at the Index level
• It is a somewhat dated classification. eg: The word corpse is used as a Head Word, where current usage would be body.

BTW, I also think this project should be a thesaurus too. --Maveric149

To quote - "A collaborative project to produce a free and complete dictionary and thesaurus in every language. " Though, like everything else in Wiktionary even that could be redefined ! Who is the final arbiter of what the Wiktionary project is about ??--Richardb 12:45, 19 Nov 2004 (UTC)

No one particular person - we tend to thrash things out in the beer parlour and come to some sort of agreement, especially on major changes such as this one. It tends to be the more regular contributors who determine what happens, for two reasons: we are around a lot to continue the discussion and put our arguments over, and we are the ones who the changes will have the greatest impact on; that said, we aren't the ones who have the final say on what happens or what doesn't - any user who suggests a solution that is popular, effective, useful and consistent with the goals of Wiktionary is likely to win the day. — Paul G 18:02, 19 Nov 2004 (UTC)

Well, there is a case for having a separate thesaurus, and we have touched on it here at some time in the past. One of the arguments reason for having a separate thesaurus is to ensure consistency of synonyms. At the moment, words with lots and lots of synonyms, especially slang terms, have in their entries variable numbers of these synonyms listed. For example, a word with n synonyms would, ideally, have n – 1 synonyms listed at each of the n entries for these words. In practice, we have somewhat less than n – 1 synonyms in the lists (in some cases, none at all). This is not only inconsistent, but unwieldy: every time some comes up with a new synonym, it has to be added to the synonym lists for the other n entries.

In the same way that rhymes have their own index (rather than listing all the rhymes for words in their entries), having a separate namespace or project would remove these shortcomings. We already have it to a very small extent with the translations - when YYY means exactly the same as XXX, and XXX is the more common term, I routinely add add "see XXX" instead of providing translations. Again, this helps to keep things consistent and more easily maintainable.

— Paul G 14:22, 19 Nov 2004 (UTC)

Folks, like yourself apparently, are already putting synonyms into wiktionary. Which is good. There's no need for a separate namespace or project. --Eean 03:46, 19 Nov 2004 (UTC)

I vote that the WikiSaurus concept is:

1. An interesting exercise.
2. Not needed.
3. Represents loss of data that belongs in the articles themselves.
4. Does not link definitions, it links words, which can be very misleading or worse, incorrect.
5. Better suited by a bot that lists synonym/antonym referees that do not link back to the referer.
6. Better suited by a bot that lists synonym/antonym references.

--Connel MacKenzie 06:00, 18 Jan 2005 (UTC)

• Response
  • Loss of Data - not one scrap of information has been lost, and plenty more added by the WikiSaurus entries. It has been reorganised. But, this is an electronic system, not a paper one, and some restructuring in this way is to be expected.--Richardb 12:15, 24 Jan 2005 (UTC)
  • A thesaurus is about linking senses/meanings, not words or definitions.
  • A Thesaurus addresses more than synonyms and antonyms, it addresses related senses, maybe nouns, adjectives, adverbs etc. Cannot see how you can get a bot to link senses/meanings at all.--Richardb 12:15, 24 Jan 2005 (UTC)

To rephrase what I said earlier (and you misunderstood):

This represents a loss of data from the article.

Now, maybe I am just misunderstanding you or this solution. Is the idea here to list wikisaurus entries in the synonyms and antonyms sections, instead of the synonyms and antonyms themselves?

(Obviously, if you clarify that, then my other points are moot. Each wikisaurus sense is linked by number to the definition sense it describes, just as synonyms now do, right?) --Connel MacKenzie 20:15, 24 Jan 2005 (UTC)

In general, I don't think I would remove the key synonyms and antonyms to put them into a WikiSaurus entry. I agree these should stay in the article. But, being a Wiki, even if I did remove them, I'm sure someone else with a different view would pu tthem back, and I wouldn't lose any sleep over that.

But, when the list of synonyms in an entry gets ridiculously long, like it was getting with "fuck" and it's synonyms, I think the list should be trimmed and replaced by a referral to the WikiSaurus entry. But again, it's a Wiki. If people don't like the trim, they'll put them back!

• The yes's give the N(N-1) argument, which is compelling, but misses the concept of the non-overlapping venn diagrams for related words. Eupedia 17:25, 2 May 2006 (UTC)
• The no's argue that the present system incorporates synonyms and related words already, but fail to solve the maintenance and completeness problem. Eupedia 17:25, 2 May 2006 (UTC)

I propose something in between, a software layer that allows:

a user to navigate the references in the syn/ant and related words sections using some kind of grouping heuristics such as latent semantic indexing (LSI), and;

an editor to immediately see the list of candidate related words (again using LSI, perhaps) and quickly select which ones he or she considers qualified to be listed in they syn, ant/ RW sections (which could be created automatically). Eupedia 17:25, 2 May 2006 (UTC)

I agree that a thesaurus in Wikitionary would be very useful. However, there are certain flaws. for instance, when I looked at the list there were about fifty synonyms for "sexual intercourse," and no words were up which I really needed. I propse, therefore, that the people thoroughly limit words which are not needed and get on with words that are actually used! I mean, it's all very interesting to know that "sexual intercourse," can be replaced with "martial knowledge," but it's not useful in actual life, is it? Another thing. Wikisaurus is currently hosted as a sub-part of Wikitionary, and makes searchers actually find the word they want, and click it. Doubtless, later it will be very frustrating for searchers to find the word they want from hundreds, no, thousands, of words.And secondly, I suggest that Wikisaurus be merged within Wikitionary to let searchers instantly find the words they want instead of having to go to Wikisaurus then double back.

All that means is that we would also have ==Synonyms== and ==Antonyms== sections in addition to the ==Translations== section. Peace.--Maveric149 The Thesaurus section is a great idea by doing it the way you have suggested, but the translation needs to be thought about. Needs a long discussion before we can decided how to do it. I suppose haveing a section is fine and if its not a perfect translation then explain why. etc -fonzy

The thesaurus stuff is more complicated than I thought since one word can have multiple definitions. Synonyms and antonyms will therefore need to be inline with the specific definitions. See dog. --mav

Hmmmm maybe something like

Dog -

1)an animal thing

2)Name for a woman

===Thesaurus===

1)Canine,

2)Bitch

Hm, that just may work. --Maveric149

Actually that might be a problem since words like dog can be both a noun and a verb (thus there would be two numbered lists). --mav

The functional subsections should probably belong to each meaning/usage separately. --Brion VIBBER

So would the dog example be a good one then? --mav

I did create a test page for an idea i have suggested before:Thesaurus:Dog.

-fonzy

"Idea" maybe there can be 2 different sections of wiktionary, wiktionary.orghttps://dictious.com/en/D:Dog for the dictionary, and wiktionary.orghttps://dictious.com/en/T:Dog for the thesaurus.

First of all a thesaurus links between definitions and not between words (as mentioned above). Secondly you should agree on the kind of relationships:

• Synonyms (BTW there are very litte real synonyms - most of them are only similar)
• Antonyms
• Upper terms/Narrower terms?
  • Class/Instance?
• or just simple links (related terms) like in the examples?

Note: A thesaurus is much more than what you know from word processors!

I suggest only two or three types of thesaurical links:

• antonyms (not:)
• similar word (syn:)
• related terms (see:)
• (maybe) upper/narrower term (UT:, NT:)

Everything more is too complicated (see w:WordNet) and leads to a wikithesaurus which is not a wiktionary anymore.

Suggestion (example from WordNet - i don't like dog/2):

==== dream ====# a series of mental images and emotions occurring during sleep

1. • syn: imagination, imaging, imagery, mental imagery
   • see: mind's eye, vision, picturing, envisioning, chimera, evocation, pretense
   • not: reality
2. a cherished desire
   • syn: desire

...

1. have a daydream; indulge in a fantasy
   • syn: daydream, woolgather, stargaze
   • see: imagine, conceive of, ideate, envisage
   • see: visualize, envision, forsee, fantasize, prefigure, think

--- nichtich

As soon as the thesaurus contains two-word or longer phrases, it becomes almost impossible to distinguish this function from that of an w:idiom dictionary.

The two functions should be considered together. See discussion in Idiom_dictionary_considerations, which could cross-fertilize well with this one.

I think it would be nice if Wiktionary were both a thesaurus and a dictionary, but the two are indeed quite different:

• a dictionary describes words, complete with pronunciation, etymology and usage.
• a thesaurus describes senses (or synsets as w:WordNet calls them: sets of synonymous words).

It seems like there should be one Wiktionary article per word (dictionary), and one Wiktionary article per sense or synset (thesaurus). Of course the dictionary page about a word would link to all its senses or synsets, as would the pages of all its synonyms. Translations to other languages seem to be more logically located at the sense (thesaurus) level rather than at the word (dictionary) level. At the thesaurus level, one could then also mention the various semantic relationships that WordNet recognizes, such as holonym, meronym etc. AxelBoldt 20:26 Jun 24, 2003 (UTC)

The > Visual Thesaurus is based on WordNet and it shows, I think, 14 different relationships including part of, member of, made of and other amazing connections that no thesaurus gives. I just discovered that it does have compound terms as well, which I had thought WordNet omitted.

Try it on the word law and then pull it around and see all the connections, to police to Scotland Yard, to Newton.

Yes, you need to have BT, NT, RT, See, and See also relationships and, yes, there is no such thing as true synonymy (even lawyer is not the same as attorney because you can be a lawyer but you need to have a client to be an attorney) but for real power and usefulness in the coming Semantic Web, you want more than a dictionary, more then a thesaurus, you want an ontology. TerrySeale --138.88.48.238 09:43, 23 May 2004 (UTC)

Can we have Wiktionary pages named like WikiSaurus:sexual intercourse where all the synonyms, anotnyms etc of Sexual Intercourse are listed (and boy are there a few!), instead of having to enter the Synonyms and Antonyms, and Relateds etc, in every word that is an Synonym of Sexual Intercourse.

Of course, there is a difficulty in who decides which words are the KEY words, but can we get around that somehow ? We can always vote for deleting unnecessary duplicates--Richardb 12:03, 19 Nov 2004 (UTC)

I've set up a couple of pages towqrds an experiment / example. See

• todger
• WikiSaurus:penis
• dick
• cock
• penis
• sexual intercourse
• WikiSaurus:sexual intercourse

Also check ] for a dynamic list of WikiSaurus articles. --Richardb 13:39, 19 Nov 2004 (UTC)

Now that I see what you mean, I agree, it makes sense. I'm not sure about the wikisaurus name though. Maybe just thes: or thesaurus: even. --Eean 01:39, 20 Nov 2004 (UTC)

Hi Richardb,

You showed me quite recently that the Project namespace is dynamic. Perhaps I would object less to the WikiSaurus concept if this were moved there, and in either the ===Synonyms=== section or even the ===Thesaurus=== section, the WikiSaurus entries were embedded, using for example, {{Project:WikiSaurus:drunk}} inside each of the terms it refers to. Can this be tried with one entry (perhaps drunk is a horrible example, having so many entries) so we can see if this would make it any easier, without losing anything from the articles themselves? --Connel MacKenzie 21:28, 4 Feb 2005 (UTC)

This essentially gets down to the Wiktionary software needing to be overhauled. I've been making a memorandum of all these ideas and their comments. I'll eventually get around to posting a Wiktionary RFC on Meta-Wiki for the future developers, if any. --Blade Hirato 09:29, 20 Nov 2004 (UTC)

Can't quite see how the software needs to be amended. Don't think an auto-generated Thesaurus is feasible, if that is what you were contemplating--Richardb 07:02, 28 Nov 2004 (UTC)

It's probably been mentioned before, but I think the wiktionary needs more structure. Editing an entire article in one big text box works for wikipedia, but maybe the wiktionary should have a bunch of clearly defined sections to edit: Name, (noun/verb/etc), pronunciation, etymology, definition, adjectives, and translations. These would be edited from single-line text boxes, or something. Then, the search box on the left could have a bunch of search boxes, one for each of the sections, and you could do really effective searches: a verb, a synonym of this, with this in it's etymology. (But it would need a complete rebuilding of the search engine, and it might take a lot of computer resources. I don't know.) Users (or anyone, with cookies) could specify what sections they like to view (ex - need a thesaurus? Just close all the sections except for synonyms.) Like this, we wouldn't need a separate thesaurus, and everyone could customize their wiktionary as they like. --64.110.250.141 18:31, 25 Feb 2005 (UTC) (Sorry, this probably wasn't the best place to mention this...)

How about implementing this using categories? This is just a bit of a brainstorm, and the words I've used as examples may be incorrect, but here's how it might work.

In the entry for stiff (as an example) you might include ] and ] which would cause it to show up under both headwords. These categories would be sub-categories within a hierarchy so that you could find information in a top-down approach too, however I would imagine the main use would be to look up a word and then look for synonyms/antonyms etc. from there.

In practice, I would suggest wrapping the category headings in a template, so you could use something like {{hw|hardness}} to produce something like this :

Use Wikisaurus to look up alternative words to stiff under the headword hardness.

I hope I've got the syntax right:

:''Use Wikisaurus to look up alternative words to '''{{PAGENAME}}''' under the headword ].''

]

The category page itself could include whatever other text is desirable (it should not be the definition of the headword however). Possibly even a concise thesaurus-like listing of the related words. Although this is duplication of the list that appears below it, this will make the entry easier to read, allow us to include words for which we don't currently have articles and let us break the list down by type of word (synonym, antonym etc). The main purpose of the categorisation if this is the case would be to lead people to the headword article, rather than the creation of a list.

Alternatively a convention of using a piped category link to apply order could be adopted, e.g. ] and ], with s meaning synonym, and a meaning antonym, and with other letters for the other word types we want to include. This would cause all synonyms to be listed together, in alphabetical order, under s, all antonyms under a, etc. It would not work so well if there are many items of each type, but if the lists are short then this could be an option.

I think there are various possible ways of implementing a thesaurus using categories which would need to be discussed, but that's just a few late-night ideas from the top of my head. Thoughts? --HappyDog 04:06, 14 Mar 2005 (UTC)

It looks pretty useful, though perhaps a bit complex to catch on much. But I do worry about the overuse of categories. Having thousands of categories can make the category system unworkable/unusable.

I tried linking WikiSaurus entries to Wiktionary:Roget Thesaurus Classification . See WikiSaurus:mad person as an example. --Richardb 16:49, 22 Mar 2005 (UTC)

Actually, I would have thought simply adding {{hw|hardness}}, or even {{hw|s hardness}} to an article is pretty easy! Much easier than having to enter the full taxonomy, as is the case in your example. Once the template is in place the casual user should be able to use it transparently. The only complication is knowing which headword to place it under, but I imagine that will be a problem with all suggestions. --HappyDog 03:33, 23 Mar 2005 (UTC)

The primary task of a thesaurus is to find the exact word you want, at least for me. A paper thesaurus is typically organized using categories, with an alphabetical index. If you know you're looking for a word like "formidable", you look it up in the aphabetical index, which refers you to a list of numbered categories, eg, "37: War". And then you'd see formidable listed under the "adjectives" section of the category, that would list the adjectives relating to war. The adjectives themselves would be grouped into sub-categories (37.1: Adjectives that describe the intensity of war; 37.2, adjectives that describe the attitudes of war; and so on, until 37.8, which would be adjectives that don't describe war, ie antonyms, such as "peaceful". 37.9 might be the first subcategory containing nouns.) The categories are numbered in such a way that related words are more likely to appear in close proximity. So, if the word I really wanted was "adamant", but the only thing I could remember was "formidable", the paper thesaurus is still a useful tool, because it's likely that adamant might appear in category 36 or 38. I'm going into this horrible level of detail because I have never seen an online thesaurus, whether on the web or integrated into a word processor, that was as useful as a paper thesaurus. Most online versions let you type in a word, and present you with either a single list of synonyms, or a list of categories that you can drill down into. That's useless. What you really want is to provide a huge list, filling the whole screen with related words. By all means, break them down into categories and subcategories, but please display more than one such category on the page. Using # anchors you can center the screen on the most relevant category and let the user scroll from there. I particularly dislike the WordNet approach of finding words: It presents you with a very small list of words and lets you "narrow" or "widen" or "find rhymes" and so on. This doesn't work for me when I'm hunting for a word -- given such a hyperstructured thing to navigate, how do you find "adamant" starting from "formidable?" Do you narrow or widen or what? It's non-intuitive. Whatever model you use to organize synonyms in Wiktionary, please look to the paper versions for their wisdom. I personally would love to see an online thesaurus organized like the paper Bartlett's/Roget's, that displayed multiple categories containing hundreds of words near each other. In this hypothetical system, clicking on a word would take you to its definition page. To get a new list of words, you'd do a new Thesaurus search. To this end, having a separate Wikisaurus seems like a good idea. However, its current implementation doesn't work for me. It's the Microsoft Word version of "enter a word and get a small list". That's not very useful for a poet.

Wikibooks (main page, about) recently put wikibooks:Body parts slang (orig wikipedia:Body parts slang) at votes for deletion. I am a Wikibooks contributor who voted for deletion after believing that "Body parts slang" was mostly thesaurus content, and noticing that Wikibooks has a thesaurus. I am also trying to change Wikibooks policy (b:Wikibooks:What is Wikibooks: "Wikibooks is not a dictionary" with reference to Wiktionary) to claim that Wikibooks is not a thesaurus. My concerns are:

1. Does wikibooks:Body parts slang seem suitable for Wiktionary? I want to avoid the scenario where stuff moves from w:Wikipedia:Articles for deletion to b:Wikibooks:Votes for deletion to Wiktionary:Requests for deletion.
2. Wikibooks might write "not a thesaurus" into policy even while Wiktionary lacks a thesaurus policy.

I hope that WikiSaurus contributors comment on this. --Kernigh 05:07, 30 October 2005 (UTC)

Someone copied it from Wikipedia to Transwiki:Body parts slang. "Wikibooks is not a thesaurus" will probably become policy soon. --Kernigh

02:51, 29 November 2005 (UTC)

If we'd like a thesaurus that is both useful and new, something that isn't merely a copycat imitation of existing paper-based thesauruses, but really capitalizes on the greater flexibility we've got in an all-electronic design, I think we want to make sure we're broadly accommodating of all the various relationships between words, not just simple synonyms and antonyms, but certainly weak/related/suggestive sorta-synonyms, and then potentially all the funky meronyms and holonyms and the 14 other Wordnet relations, too. But of course an important challenge will be make sure that the funky obscure relations don't obscure the familiar ones, to make sure that an unsophisticated reader just looking for simple synonyms and antonyms won't be overwhelmed.

But there was an exremely important comment made a year ago by 216.38.149.153 pointing out that one big use (me, I'd even say the big use) of a thesaurus is finding that perfect word that you can't quite remember but that you know when you see it. But as 216.38.149.153 pointed out, a too-logical, too-structured, too-mechanical system doesn't do that for you:

"I have never seen an online thesaurus that was as useful as a paper thesaurus. Most online versions let you type in a word, and present you with either a single list of synonyms, or a list of categories that you can drill down into. That's useless... It's the Microsoft Word version of "enter a word and get a small list". That's not very useful for a poet."

So we've got to capture those "near" and "similar" and "related" and "you might also be thinking of" relations as well. There are several ways to approach this, I think:

1. Make sure we've got "similar to" and "related to" and "you might also be thinking of" as relations.
2. Make sure we're integrated with a Roget-like, concept-based, sense-and-meaning classification hierarchy, which is somehow also browsable.
3. Do implement sets or lists of near-synonymous words (perhaps precisely what Wordnet calls "synsets"), but then work on ways of linking and grouping them.

With respect to #3, Roget does a lot of this. For example, here's an entry picked somewhat at random:

#740. Lenity.-- N. lenity, lenience, leniency; moderation &c. 174; tolerance, toleration; mildness, gentleness; favor, indulgence, indulgency; clemency, mercy, forbearance, quarter; compassion &c. 914.

Notice that there are a bunch of little lists there, with semicolons between them. More importantly, notice the cross-references to other related entries. When I use a paper-based thesaurus, I often find myself chasing those all over the place. And they're obviously much easier to create and maintain -- not to mention to use! -- in a hyperlinked electronic format like this.

—Scs 05:13, 4 May 2006 (UTC)

Here's a much more extensive example from Roget, showing a three-level internal hierarchy (comma, semicolon, full stop):

#837. Dejection. --

...

Adj. cheerless, joyless, spiritless; uncheerful, uncheery; unlively; unhappy &c. 828; melancholy, dismal, somber, dark, gloomy, triste, clouded, murky, lowering, frowning, lugubrious, funereal, mournful, lamentable, dreadful.

dreary, flat; dull, dull as a beetle, dull as ditchwater; depressing &c. v.

"melancholy as a gib cat"; oppressed with melancholy, a prey to melancholy; downcast, downhearted; down in the mouth, down in one's luck; heavy-hearted; in the dumps, down in the dumps, in the suds, in the sulks, in the doldrums; in doleful dumps, in bad humor; sullen; mumpish, dumpish, mopish, moping; moody, glum; sulky &c. (discontented) 832; out of sorts, out of humor, out of heart, out of spirits; ill at ease, low spirited, in low spirits, a cup too low; weary &c. 841; discouraged, disheartened; desponding; chapfallen, chopfallen, jaw fallen, crest fallen.

sad, pensive, penseroso, tristful; dolesome, doleful; woebegone; lacrymose, lachrymose, in tears, melancholic, hypped, hypochondriacal, bilious, jaundiced, atrabilious, saturnine, splenetic; lackadaisical.

serious, sedate, staid, stayed; grave as a judge, grave as an undertaker, grave as a mustard pot; sober, sober as a judge, solemn, demure; grim; grim-faced, grim-visaged; rueful, wan, long-faced.

disconsolate; unconsolable, inconsolable; forlorn, comfortless, desolate, desole, sick at heart; soul sick, heart sick; au desespoir; in despair &c. 859; lost.

overcome; broken down, borne down, bowed down; heartstricken &c (mental suffering) 828; cut up, dashed, sunk; unnerved, unmanned; down fallen, downtrodden; broken-hearted; careworn.

Another sort of "relation" which might be captured, not applicable to all words but useful and interesting for those it does apply to, is position in and connection with neighbors along a spectrum. I haven't seen a "-nym" term for this. For example, two such spectra are freezing/cold/chilly/cool/neutral/tepid/warm/hot/scalding/incandescent and terrible/bad/lousy/poor/neutral/fair/middling/good/excellent. —Scs 05:53, 4 May 2006 (UTC)

That sounds like an excellent extension of WikiSaurus. This sort of extension to some extent validates using a new name, rather than just using the title thesaurus, as we can go much further than a usual Thesaurus, just as Wiktionary goes well beyond a normal dictionary. I suggest "creating" a -nym word for this, to help us describe what is meant. If and when someone finds a real word to mean this, we can replace the occurrences of your made-up word with the real word. I've got some suggestions, but will leave to you the privilege of being creative about the word.--Richardb 03:35, 11 May 2006 (UTC)

Thesaurus design is quite complicated. Related Topics are:

• w:Thesaurus
• w:WordNet
• w:Semantic_network

Whilst WikiSaurus is still as raw as can be, and controversial, has it already grown in scope above that of what might be termed "Simple" Thesaurus ? Should we call a close to this discussion of "Thesaurus considerations", and have a place (or group of places) discussing the scope, ongoing development and improvement of WikiSaurus ? In no-way would this pre-suppose how WikiSaurus should be implemented (the present method is still a somewhat controversial trial exploring the possibilities). I'm just looking at trying to rationalise, centralise, focus the discussion of what we are increasingly calling "The WikiSaurus", as opposed to talking about a Thesaurus and the WikiSaurus as two seprate conversations.--Richardb 03:54, 11 May 2006 (UTC)