Content Partnerships Hub/Helpdesk/Malayalam lexemes/Lexeme guide

	Content Partnerships Hub
	Improving the Wikimedia movement’s work with content partners

Beginner guide to Wikidata lexemes

About this guide

This guide outlines the most important things to know for newcomers to the Lexicographical data namespace in Wikidata. It covers the basics of how lexemes are structured, as well some of the most popular tools that make editing them easier.

It is meant to function as a Quick start guide. More in-depth explanations can be found in the Wikidata documentation, as well as in the documentation of the specific tools.

Lexeme overview

Basics of Lexeme structure

Lexemes in Wikidata are designed to represent words or phrases in a structured and consistent way.

Every lexeme is made up of three main parts:

The lexeme itself:

a lemma – the base form of the word (e.g. cat)
a language – which language the word belongs to (e.g. English (Q1860))
a lexical category – part of speech (e.g. noun (Q1084), verb (Q24905), adjective (Q34698))

Every lexeme has a unique identifier, beginning with the letter L, e.g. L7.

The forms

Variants of the lemma that occur due to inflection.

🟩 Example: for the lexeme cat the plural cats is stored as a form.

🟩 Example: for the lexeme teach, the inflections teaches, taught and teaching are stored as forms.

Forms are annotated with grammatical features such as number, gender, tense, mood, or case.

🟩 Example: teaches is annotated with third person, singular, simple present.

ℹ️ Note: every grammatical feature is a Wikidata item. See for example: plural (Q146786), supine (Q548470), attributive (Q4818723).

The senses

The meaning of the word, stored as a short dictionary-style definition.

A lexeme can have multiple senses.

A sense can be expressed in multiple languages. It is technically not obligatory to add a sense in the language of the lexeme, though of course it is helpful.

🟩 Example: teach has the sense to pass on knowledge.

A sense can be linked to a Wikidata item corresponding to the meaning of the lexeme, using the property item for this sense. This is mostly relevant for nouns, since they will often correspond to specific concepts that also exist as Wikidata items.

🟩 Example: the lexeme geography has the sense science of Earth's surface and societies on it which is linked to the Wikidata item geography (Q1071).

ℹ️ Note: we link lexeme senses to items, which implies a sense has to be added first.

This link is important because it anchors the short definition of a sense to a precise concept in Wikidata's knowledge graph, ensuring that the meaning is unambiguous and machine-readable. By using this property, senses become interoperable with the rest of Wikidata's data: they can be queried, reused in applications like Abstract Wikipedia, and connected to other resources that already link to the same item.

Item for this sense and translations

When lexemes in different languages have senses that are linked to the same Wikidata item, it indicates that they represent the same concept.

Example: geography, جُغْرَافِيَا and geografi are three different lexemes. Each of them has a sense that links to the item geography (Q1071).

Thanks to this, we know that they express the same concept, science of Earth's surface and societies on it, in different languages.

Statements on lexemes

Just like items, lexemes, their forms, and their senses can all have statements.

These statements add extra information that makes the data richer, easier to query, and more useful.

On the lexeme itself

Lexemes can have properties that describe general information about the word as a whole. For example:

Usage example – provides a real usage example for the lexeme.

The usage example can be accompanied by the qualifiers subject sense (P6072) and subject lexeme form (P5830) to indicate which form and sense of the lexeme it demonstrates.

If the usage example comes from an authentic source (like a website or a book), it should also be accompanied by a reference.

Combines lexemes – indicates the lexemes combined in this lexeme

Many languages, such as Swedish, German and Norwegian, make a heavy use of compound words. Using this property, we can indicate which lexemes the given lexeme is a compound of. Use the qualifier {{{}}} to indicate the order of the combined lexemes; since it does not matter in which order you add them.

Language-specific statements

Depending on the specific language, there might be specific statements the community has agreed on to add.

🟩 Example: in German, nouns have genders. Thus, Schule has the property grammatical gender (P5185) with the value feminine (Q1775415).

🟩 Example: in Swedish, verbs belong to different conjugation classes. Thus, springa has the property conjugation class (P5186) with the value Germanic strong conjugation class 3a (Q56649335).

External identifiers

You can use external identifier properties to link to the corresponding entry in an external source, such as an online dictionary.

Example: binda has the property Svenska Akademiens ordlista ID (P11838) with the value 0262821.

In Ordia, you can see the statistics of external identifiers used in lexemes. Among the most used ones are the German Duden lexeme ID (P8376) and DWDS lemma ID (P9940).

Questions people often ask

How do I search for a lexeme?

In Wikidata, you can use the search bar and preface your search with L:, e.g. L:seventeen. Note that autofill will not work, so you won't see any results until you actually execute the search.

Using the search function in Ordia is more convenient, as it's limited to the lexicographical namespace by default.

How do I know which lexemes need to be created?

The Text to Lexemes tool in Ordia is very useful for discovering missing lexemes. You can paste any text there, like a Wikipedia article of your interest. If you want to work in a more structured way, you can try and find a frequency list of your language and use it with Text to Lexemes.

Wikidata:Lexicographical coverage is an automatically updated report of the most common word forms in a given language that are missing in the lexeme namespace, based on corpus data (mostly Wikipedia).

How do I know I'm not creating a duplicate lexeme?

If you create a lexeme directly in Wikidata, using Special:NewLexeme, you won't know if you've accidentally created a duplicate.

That's why many editors prefer using exclusively Lexeme Forms, as it will warn you about potential duplicates.

How do I know the conventions and rules for my specific language?

The community of each language decided on their own best practices and guidelines. They are collected in this category on Wikidata. For example, here are the guidelines for Swedish.

If you language is missing, or the documentation is sparse, be bold and help out!

What's the difference between lexemes and Wiktionary?

Wikidata and Wiktionary work on very different principles. Wikidata allows us to store the data about language in a structured and machine-readable way, while Wiktionary is much more flexible, allowing for more detailed definitions, descriptions of etymology and so on.

Many Wiktionary editors prefer to edit only their project, and the other way round.

You can read more about the relationship between the two projects in Wikidata:Wiktionary.

Can I add etymology, usage notes, or long explanations?

Simple etymological statements, that can be expressed using linked data, are possible to add. The property derived from lexeme (P5191), for example, can be used to indicate where a lexeme is derived from; first attested from (P6684) indicates where a word was first attested, etc.

However, for more nuanced explanations, usage notes and the like, Wiktionary is more appropriate.

What lexemes can I add?

Have a look at Wikidata:Lexicographical_data/Notability which outlines the notability criteria for lexemes. In short, you should be able to prove that the lexeme is present in serious and publicly available sources, or fulfills a structural need.

Can I use QuickStatements or OpenRefine to work on lexemes on a larger scale?

QuickStatements can be used to add statements to existing lexemes, their senses and forms.

OpenRefine does not support the lexicographical namespace (as of August 2025).

Useful tools

Working with lexicographical data in Wikidata can be made much easier and more efficient by using a variety of community-built tools. While it is always possible to create and edit lexemes directly through the standard Wikidata interface, specialized tools help with tasks like generating whole paradigms of forms, browsing lexemes by language or category, or performing bulk edits. Some tools also let you customize your editing environment with gadgets and scripts so that frequently used actions are quicker and more accessible. This chapter introduces the most important tools for beginners, explains what they are useful for, and provides links so you can start experimenting with them right away.

Ordia

➡️ Go to the tool

Ordia is a useful tool for exploring and browsing lexemes. It can also assist you in finding missing lexemes based on texts.

Language statistics

Language statistics provides a quick overview of the languages represented in Wikidata's lexicographical data. You can see how many lexemes a language has, as well as some correlations, such as the relationship between the number of lexemes and the number of senses for a given language.

Browse lexemes by language

🟩 Example: Estonian lexemes, with statistics of lexical categories.

View individual lexemes

🟩 Example: folk in Swedish. This is an alternative to viewing the lexeme in Wikidata. It provides several extra functionalities compared to Wikidata, as it pulls and displays data from lexemes that link to the displayed lexeme. For example, if the displayed lexeme is part of any compounds, this will be displayed directly, without you having to write a SPARQL query to find those compounds.

Search across lexemes

The search function in Ordia can feel more user-friendly than that in Wikidata.

Text to lexemes

Text to lexemes allows you to process any text, e.g. a chunk of a Wikipedia article, to find whether the words in it have Wikidata lexemes. If a lexeme exists, you can also see if it has any senses.

If a lexeme does not exist, you can create it directly from Ordia.

Why is it useful?

You can quickly check which words in a text are already represented as lexemes
You can spot missing lexemes: words in your text that do not yet exist in Wikidata's lexicographical data
You can find gaps in specific languages, or prepare lists of lexemes to create

Lexeme Forms

➡️ Go to the tool

🗒️ Read documentation

Lexeme Forms is a tool that makes creating lexemes much faster and easier than using Wikidata directly. Most prolific lexeme creators rely on it.

How it works

The main page of the tool lists available languages and lexeme types (templates). Each template cues you to enter different forms of the word.

🟩 Example: for Swedish, we can select between different genders of nouns, as well as nouns and different types of adjectives (comparable and not comparable).

Why is it useful?

There are several aspects of the tool that make it useful for prolific contributors:

The tool checks whether any duplicates (lexemes of the same lexical category in the given language) exist. Wikidata's interface does not do that. Note that while the tool will discourage you from creating a duplicate lexeme, you can still confirm you really want to do so (homonyms exist!)
The tool does the work of adding grammatical tags (such as active (Q1317831), present tense (Q192613)...) to each of the forms. Those are defined in the templates and added automatically. This means that you don't have to remember which tags go with which form! This makes the tool particularly useful for introducing beginners to lexeme editing.

How are the templates created? What if a template I need does not exist?

The templates are created by the community (by you!) In the documentation, you will find information on how to add a new language or template.

Lexeme Forms gadget

In your user preferences in Wikidata, you can enable the Lexeme Forms gadget. Then, when looking at a lexeme in Wikidata, if a matching template corresponding to its language and lexical category exists in Lexeme Forms, a link to the tool will be displayed in the sidebar. That's useful for spotting and adding missing forms to lexemes.

Bulk mode

Every template in Lexeme Forms has a bulk mode, where you can create a larger number of lexemes using the same template, using a QuickStatements-like syntax. Duplicate lexemes will not be created. Of course, you should double and triple check your input beforehand!

Gadgets and user scripts

There are some smaller gadgets and user scripts that make browsing and working with lexemes easier.

LexemeEntitySuggester.js

This script adds various entity suggestions on lexeme pages. Useful e.g. when adding combines lexemes (P5238) statements, as it will search for potentially matching lexemes instead of you having to search for them, which makes the process faster.

To use it, add the following line to your common.js:

mw.loader.load("//www.wikidata.org/w/index.php?title=User:Nikki/LexemeEntitySuggester.js&action=raw&ctype=text/javascript");

LexemeTranslations.js

This script shows translations for a lexeme, automatically derived from item for this sense (P5137) statements. Note that if the item for this sense (P5137) was added recently, it can take a couple minutes for it to be updated.

To use it, add the following line to your common.js:

mw.loader.load("//www.wikidata.org/w/index.php?title=User:Nikki/LexemeTranslations.js&action=raw&ctype=text/javascript");

Queries

This is a collection of lexeme-specific SPARQL queries. They can be useful for both exploring a specific language and identifying lexemes that need attention (such as adding senses or usage examples).

They can be modified to search for lexemes in a language of your choice.

Full glossary

All lexemes in a language that have a gloss (sense) in that language.

SELECT ?l ?lemma ?posLabel ?gloss WHERE {
  ?l rdf:type ontolex:LexicalEntry;
   dct:language wd:Q809; # change language here
   wikibase:lexicalCategory ?pos;
   wikibase:lemma ?lemma.
  ?l ontolex:sense ?sense.
  ?sense skos:definition ?gloss.
  FILTER((LANG(?gloss)) = "pl") # change language here
  SERVICE wikibase:label { bd:serviceParam wikibase:language "en". }
}
ORDER BY (?lemma)

Click to try !

Lexemes without sense

All lexemes in a language that do not have a gloss (sense) in that language.

SELECT ?l ?lemma ?posLabel WHERE {
  ?l rdf:type ontolex:LexicalEntry;
    dct:language wd:Q9027; # change language here
    wikibase:lexicalCategory ?pos;
    wikibase:lemma ?lemma.
  MINUS { ?l ontolex:sense ?sense. }
  SERVICE wikibase:label { bd:serviceParam wikibase:language "sv". }
}

Click to try !

Lexemes without a main level statement

In this example, we're looking for Swedish nouns that are missing a grammatical gender (P5185) statement, which is obligatory for Swedish lexemes.

SELECT ?l ?lemma WHERE {
   ?l a ontolex:LexicalEntry ;
        dct:language wd:Q9027 ; # language = Swedish
        wikibase:lexicalCategory wd:Q1084; # is a noun
        wikibase:lemma ?lemma .
  MINUS { ?l wdt:P5185 []. } # without grammatical gender
}
ORDER BY ?lemma

Click to try !

Lexemes without forms

Lexemes with 0 forms. Every lexeme should have at least one form. We also display the lexical category of each lexeme.

SELECT ?l ?lemma ?posLabel WHERE {
  ?l rdf:type ontolex:LexicalEntry;
    dct:language wd:Q7918; # change language here
    wikibase:lemma ?lemma;
    wikibase:lexicalCategory ?pos.
  FILTER(NOT EXISTS { ?l ontolex:lexicalForm ?form. })
  SERVICE wikibase:label { bd:serviceParam wikibase:language "en". }
}
ORDER BY (?posLabel)

Click to try !

Lexemes with a certain number of forms

Sometimes you want to find lexemes that have a certain number of forms. For example, if every noun in your language is expected to have at least 4 forms, but a lexeme only has 1, it's a sign that something is missing.

SELECT ?l ?lemma ?posLabel (COUNT(?form) AS ?formcount) WHERE {
  ?l rdf:type ontolex:LexicalEntry;
    dct:language wd:Q9027; # change language here
    wikibase:lemma ?lemma;
    wikibase:lexicalCategory ?pos;
    ontolex:lexicalForm ?form.
  SERVICE wikibase:label { bd:serviceParam wikibase:language "en". }
}
GROUP BY ?l ?lemma ?posLabel
HAVING (?formcount = 1 )
ORDER BY (?posLabel)

Click to try !

Lexemes with forms that do not have grammatical feature tags

Forms without grammatical feature tags are often a sign that something is missing. In this example, we're looking specifically at verb forms without feature tags.

SELECT ?l ?lemma ?form WHERE {
  hint:Query hint:optimizer "None".
  ?l dct:language wd:Q9027; # change language here
    wikibase:lemma ?lemma;
    wikibase:lexicalCategory wd:Q24905; # verb
    ontolex:lexicalForm ?f.
  MINUS { ?f wikibase:grammaticalFeature []. }
  ?f ontolex:representation ?form.
}

Click to try !

Lexemes that don't have a specific form

Some lexemes are always expected to have a specific form. For example, in many languages, verbs are expected to have an infinitive (Q179230) form. Here we're looking for Swedish verb that do not have it.

SELECT DISTINCT ?l ?lemma WHERE {
  ?l rdf:type ontolex:LexicalEntry;
    dct:language wd:Q9027; # change language here
    wikibase:lemma ?lemma;
    wikibase:lexicalCategory wd:Q24905. # verb
  MINUS { ?l ontolex:lexicalForm ?form.
    ?form wikibase:grammaticalFeature wd:Q179230. # infinitive
        }
  SERVICE wikibase:label { bd:serviceParam wikibase:language "en". }
}

Click to try !

Lexemes that have a sense, but no examples

Helpful for identifying lexemes that could benefit from a usage example (P5831).

SELECT ?l ?lemma WHERE {
  ?l rdf:type ontolex:LexicalEntry;
    dct:language wd:Q9035; # change language here
    wikibase:lemma ?lemma;
    ontolex:sense ?sense.
  FILTER(NOT EXISTS { ?l wdt:P5831 ?example. })
}
ORDER BY (?lemma)

Click to try !

Lexemes that have a usage example, but no senses

If there's a usage example, it's probably easier to write a definition for the word.

SELECT ?l ?lemma ?example WHERE {
  ?l a ontolex:LexicalEntry ;
       dct:language wd:Q809 ; # change language here
       wikibase:lemma ?lemma .
  ?l wdt:P5831 ?example.
  FILTER NOT EXISTS {?l ontolex:sense ?sense }
}
ORDER BY ?lemma

Click to try !