This is the main project page of Wiktionary Thesaurus, a Wiktionary subproject and a wiki namespace aiming at creating a thesaurus, a dictionary of synonyms, antonyms, and further semantically related terms such as hyponyms, hypernyms, meronyms, and holonyms. The project was formerly called Wikisaurus.

New to the thesaurus?

Please contribute your own thesaurus entries, or add to the existing thesaurus entries.

Purpose

Chunked purpose
  • Purpose:
    • To help people find words that they
      • can't recall or
      • don't know
    • To help people explore the network of words
  • Semantic relations:
    • Synonymy – same or similar meaning
    • Antonymy – opposite meaning
    • Hyponymy – narrower meaning, subclass
    • Hypernymy – broader meaning, superclass
    • Being an instance – being an example of the class, set membership
    • Meronymy – part, such as wheel of a car
    • Holonymy – whole, such as car of a wheel
  • Users:
    • Writers
    • Managers
    • Contributors to wikis
    • Bloggers
    • Writers of love letters
    • Journal writers

The purpose of Wiktionary Thesaurus is to serve the role of an electronic thesaurus—a dictionary of synonyms, near-synonyms, antonyms, and near-antonyms, and also of other semantically related terms such as hyponyms, hypernyms, meronyms, and holonyms.

The purpose of such a thesaurus in general is mainly to help anyone who writes for living or fun—writers, managers, contributors to wikis, bloggers, and writers of love letters—to find words they don't recall or even know when they recall words that are semantically related to the sought word. In general, anyone to whom the choice of words matters can benefit from a thesaurus, especially one linked to a dictionary providing the definitions.

Added value of Wiktionary Thesaurus is its Wiktionary integration—it links to and is linked from Wiktionary.

Browse

To start browsing the thesaurus, you can start at the root, Thesaurus:entity, and proceed from there through the hyponymic network to high genera such as Thesaurus:person, Thesaurus:organism, Thesaurus:animal, Thesaurus:plant, and Thesaurus:artifact and further down. You can also browse by topical thesaurus category such as Category:Thesaurus:Geography (Thesaurus:forest), Category:Thesaurus:Personality (Thesaurus:humble) or Category:Thesaurus:Appearance (Thesaurus:beautiful).

Model

The thesaurus is organized primarily on the model of WordNet. That is, the key organizing principles are the relations of hyponymy (subclass) and hypernymy (superclass), and to a lesser extent meronymy (part of a whole) and holonymy (whole of the part). See also Wiktionary:Semantic relations. The design of Roget's 1911 thesaurus is somewhat similar in that it does not restrict the entries to lists of synonyms and antonyms; however, Roget's thesaurus does not use WordNet relations. The design of Oxford English Dictionary thesaurus is somewhat similar in that it has a hierarchically organized thesaurus. However, its subordination relation is not a strict hyponymy but is in part thematic. By contrast, the thesaurus of Merriam-Webster has synonyms, antonyms and words "related". Editors who want to create thesaurus entries that are primarily for lists of synonyms can do so without worrying about the other relations, but keep in mind that there should be only a single thesaurus entry for a synonym set, thereby avoiding duplication.

One sense per entry

Each entry should ideally have a single sense. Nonetheless, the format supports multiple senses for the cases where this seems to be the best option. It is usually possible to pick different headwords for different senses. Each entry should ideally stand for a semantic object; the headword should be in part an accident. The point is not to list all senses of the headword. Thus, there can be a single sense in Thesaurus:sound, and another sense of "sound" is covered at Thesaurus:inlet. If it becomes impossible to keep finding dedicated headwords, we may resort to disambiguating naming like "rich (wealthy)" or "rich, wealthy". Sometimes, the headword becomes less ambiguous by using a phrase: there is Thesaurus:English language. WordNet seems to do fine mostly by the comma convention.

To use the entry headword as a basis for covering all the senses of the headword would lead to a duplication of synonym rings covered in other headwords, at odds with the duplication-avoidance rationale for the thesaurus. Thus, there is no point in duplicating Thesaurus:spicy in Thesaurus:hot. By contrast, having a sense for an adjective and a sense for a noun in Thesaurus:German is a different use case and makes a little bit more sense, being caused by a lack of suitable English disambiguating headword unless one opts for "German person" and "German language" as headwords.

Multilingualism

English Wiktionary Thesaurus shall contain entries for other languages than English[1].

Category:Thesaurus entries by language features these entries. Categorization is done by {{ws sense}}.

Historically, there was no agreed-upon naming scheme for non-English thesaurus entries. The following conventions existed:

  1. No language code, native headword. For example: Thesaurus:صار (Arabic), Thesaurus:yaxşı (Azerbaijani), Thesaurus:chat (French), Thesaurus:god (English and Danish). Entries in different languages with identically spelled titles are placed on the same Thesaurus page, exactly as we do for ordinary dictionary entries.
  2. Language code, native headword. For example: Thesaurus:fr:embêter (French), Thesaurus:sga:ar (Old Irish), Thesaurus:non:sverð (Old Norse).
  3. Language code, English headword. For example: Thesaurus:da:beautiful (Danish), Thesaurus:ar:become happy (Arabic), Thesaurus:sound/fi (Finnish).

By October 2022, convention A was followed by 98.8% of non-English thesaurus entries. In November 2023, all remaining thesaurus entries using convention B and C were standardised to convention A.

Convention A carries certain disadvantages. One is that the automatic display of the [⇒ thesaurus] link next to synonyms and other terms can be triggered in situations where it is not relevant. For example, Thesaurus:yes lists "ar" as a synonym, but the [⇒ thesaurus] link next to "ar" points to a page containing synonyms for an unrelated term in a different language. Another disadvantage is that the scheme is unique to English Wiktionary. Having Thesaurus entries for multiple languages on a single page causes problems for interwiki linking to other-language Wiktionary editions, such as French Wiktionnaire.

Topical categorization is an unsolved problem: there is Category:Thesaurus:Geography, but no language-specific one. We could create "Category:Thesaurus:en:Geography", "Category:Thesaurus:es:Geography", etc., on the model of mainspace topical categories.[2]

Discussions:

Formatting

Formatting is specified and discussed at:

Example entries:

Inclusion

As for entry headwords, they must be attested. Not all mainspace entries should have their own thesaurus entry: the point of thesaurus is in part to prevent duplication of lists. The headwords can sometimes be sum of parts if deemed preferable, as in Thesaurus:beautiful person.

As for list items, all items in lists of synonyms, antonyms, hyponyms, etc. on Thesaurus pages are required to be attested, using the same attestation criteria as the mainspace.[3] There is no requirement that they must be more than sum of parts. Roget's Thesaurus did include many sum-of-parts phrases.

Semantic relations

See also Wiktionary:Semantic relations.

If you want to create synonym-only entries, you do not need to worry about the other relationships all that much. This is especially true of adjectives. For nouns, it often pays off to figure out a good node in the hyponymic (subclass/superclass) network.

Synonyms and antonyms

Synonyms are terms with the same or very similar meaning. Register (informal, vulgar, etc.) does not impact synonymy. Examples: Thesaurus:wise, Thesaurus:drunk. Some putative synonyms are better classified as hyponyms.

Antonyms are terms with opposite meaning. Antonyms are sometimes concentrated in an opposite thesaurus entry. Example: Thesaurus:drunk.

Hypernyms and hyponyms

Hypernyms are terms with broader meaning, capturing a superclass relationship: X is a hypernym of Y if each Y is also an instance of X. Example: Thesaurus:bird.

Hyponyms are terms with narrower meaning, capturing a subclass relationship: X is a hyponym of Y is each X is an instance of Y. Examples: Thesaurus:drunk, Thesaurus:bird. In many entries, hyponyms can be listed only up to a point, to some nesting level. For instance, it makes no sense to list all hyponyms in Thesaurus:person; by contrast, listing all hyponyms in Thesaurus:relative or Thesaurus:musician seems fine.

Holonyms and meronyms

Holonyms are terms for wholes containing parts: X is a holonym of Y if Y is part of X. Example: Thesaurus:relative.

Meronyms are terms for parts of wholes: X is a meronym of Y if X is part of Y. Example: Thesaurus:aircraft.

Classes and instances

X is a class of Y if Y is an instance of X, different from hypernyms. Example: Thesaurus:Ecuador.

Instances are opposite of classes, different from hyponyms. Example: Thesaurus:country.

Coordinate terms and troponyms

Coordinate terms, also known as cohyponyms, are mostly unused in the thesaurus since it duplicates hyponymic structures from other entries.

Troponyms are unused: use hyponyms and hypernyms for verbs as well.

Various

The section "Various" is intended to capture other interesting relations, to broaden the navigation network beyond specifically defined relations. It supports creativity, but may lead to disagreements between editors since there is no set of specific rules governing the section.

Example entries:

  • Thesaurus:number: has all sorts of terms relating to numbers that are not hyponyms or instances.
  • Thesaurus:size: has adjectives for size and these do not fit hyponymy or instance-of relationships.
  • Thesaurus:aircraft: has people on board, who are strictly speaking not meronyms.

Minimum item count

A putative thesaurus entry with 2–5 items can probably be comfortably handled by the mainspace synonym lists, and may be not worth an entry. However, there is no agreed on rigid rule for this. The thesaurus most pays off when the item counts are larger.

There is usually no need to create "leaf node" entries for 1 or 2 synonyms and 1 hypernym. Such items are sufficiently covered in the hypernym entries and in the mainspace. Thus, there is Thesaurus:lake but no Thesaurus:pond.

Templates

Lists of templates:

Templates:

Template Example Note
  • {{ws|en|beer|An alcoholic fermented malt drink.}}
  • {{ws|en|beer|*}}
  • {{ws|en|beer}}
The second entry is a lazy one. The third one shows no tooltip, but also does not make it clear that one is missing.
To be put around a list of {{ws}} entries. Currently formats the list as a 3-column one.
Entered at the very top of the entry. When without parameter, determines the headword automatically.
{{R:Roget 1911|beauty}} beauty” in Roget's Thesaurus, T. Y. Crowell Co., 1911.
{{ws sense|en|glad; in a good mood}} Sense: glad; in a good mood Used after the third level heading for the part of speech.

Mainspace

Linking from mainspace to Thesaurus entries:

  • Links to thesaurus entries can be added to the "Synonyms" section (or "Hyponyms", "Antonyms", etc. where appropriate) using the template {{seeSynonyms}} (which displays something like: See Thesaurus:error), or using conventional wikitext syntax (''See also'' [[Thesaurus:error]]).
  • The {{synonyms}} template, used to render per-sense synonyms directly beneath definitions, accepts Thesaurus: links, which should be placed after any specific synonyms (e.g. {{synonyms|en|goof|blunder|Thesaurus:error}})
  • Especially for Thesaurus entries featuring mostly synonyms, it is good to add a link to the Thesaurus entry from all the mainspace entries for the synonyms, so that the user knows that there is a Thesaurus entry when visiting the mainspace.

Wikidata

Wikidata with its subclass and instance of relationships does some of the job of the thesaurus, and is hugely more complete. However, it is not suited for extensive synonym lists and it does not make it convenient to browse hyponymic networks, only supporting easy navigation from an item to its superclass. Some of its subclass stuctures seem needlessly complex and overengineered.

Roget-MICRA thesaurus

Roget's 1911 thesaurus with MICRA supplementation is available here:

The appendix has a search box and conveniently features links to mainspace.

A search box for convenience:

Moby Thesaurus II

Moby Thesaurus II is available here:

The appendix has a search box and conveniently features links to mainspace.

A search box for convenience:

Identity

The current title of the project is "Thesaurus" and "Wiktionary Thesaurus". Before mid-2017, it was "Wikisaurus"[4] Alternatives considered include "Wikithesaurus". In the past, WikiSaurus spelling with capital 'S' must have existed at some point.

Online thesauri

Public domain

Free as in "freedom"

Proprietary

Other

  • None listed.

Statistics

Statistics about the thesaurus entries, as of Oct 2022:

  • Entries: 4,833
  • English entries: 2,487
  • Chinese entries: 1,900
  • Other-language entries: 446
  • Entries containing colon (:) in title: 29

Page views

Anatomy entries get a fair amount of page views, as is expected. But they are not alone; other entries with non-trivial page views include Thesaurus:pros and cons and Thesaurus:child.[3]

Recent changes

Shortcuts

See also

Subpages

Highlighted subpages:

Project subpages:

To do

Things to do:

All entries

Lists of all Thesaurus entries:

Discussion

Discussions about Thesaurus are scattered across various pages. In the future, they should better take place in Beer Parlour, a general policy discussion room.

Pages with discussions:

For more discussions, see #Beer parlour.

Beer parlour

Discussions about Thesaurus at Beer parlour:

The following list is highly incomplete.

See also search for "Wikisaurus" in the archives of Beer parlour.

Index

An index to this page:

References