A lexeme is an abstract unit of vocabulary considered independently of its particular grammatical forms. In linguistics, the English forms write, writes, wrote, written, and writing can all represent the same lexeme, conventionally written WRITE. The concept distinguishes a vocabulary item from the forms it takes in speech or writing, providing a basic unit for studying morphology, lexical semantics, and dictionary organization.
Lexemes, word forms, and occurrences
The everyday term word can refer to several different things. The concept of a lexeme separates an abstract vocabulary unit from its realizations and from individual occurrences in a text.
- A lexeme is the vocabulary unit, such as DOG.
- A word form is a particular realization of that unit, such as dog or dogs.
- A token is an individual occurrence of a form in speech or writing.
In the sentence “The dogs chased other dogs,” dogs occurs twice. These are two tokens of the same word form, both representing the lexeme DOG.
Lexemes are often written in small capitals, or in ordinary capitals when small capitals are unavailable. This is a notational convention, not a claim about how the corresponding words are spelled.
A lexeme is more than a collection of spellings. It normally has associated grammatical properties and a conventional meaning or range of related meanings. Conversely, identical spelling does not establish that two expressions represent the same lexeme.
Inflection and paradigms
Inflection produces grammatical forms of a lexeme rather than, in the usual analysis, creating new lexemes. In English, WALK has forms including walk, walks, walked, and walking. Their differences express grammatical information such as tense, agreement, or participial status while preserving the identity of the vocabulary item.
The organized set of grammatical forms associated with a lexeme is its inflectional paradigm. A paradigm can be represented as a mapping from combinations of grammatical features to word forms. This mapping need not be one-to-one:
- Several grammatical feature combinations can share a form. English walked, for example, serves as both a past-tense form and a past participle. Such identity of forms is called syncretism.
- One lexeme can have alternative realizations for the same grammatical function, as with learned and learnt.
- Forms belonging to one lexeme can differ substantially in shape. The relationship between go and went exemplifies suppletion.
A lexeme therefore cannot always be identified by finding a string of sounds or letters shared by all its forms. Grammatical relationships and conventional lexical identity also matter.
Lexemes and related units
Lemmas
A lemma is a conventional citation form used to name or index a lexeme. English dictionaries generally cite verbs under a base form such as write and count nouns under a singular form such as dog.
The lexeme is the abstract vocabulary unit; the lemma is the form selected to represent it. Although the terms are sometimes used loosely as equivalents, they are not identical. A lemma can itself occur as an ordinary word form, and the same citation spelling can identify more than one lexical entry.
Morphemes, roots, and stems
A morpheme is a minimal unit of grammatical or lexical significance. A lexeme need not consist of only one morpheme. UNHAPPINESS, for example, is a single lexeme with a complex internal structure.
A root is the central morphological element within a word, while a stem is a form to which inflectional material can attach. These are units of morphological structure rather than alternative names for a lexeme. The same root can contribute to several lexemes, and a lexeme may have more than one stem across its paradigm.
Dictionary entries
A lexeme is also not necessarily identical to a dictionary entry. Dictionaries make editorial decisions about whether to combine or separate meanings, grammatical categories, spelling variants, and historical developments. Their headwords provide evidence about lexical organization, but dictionary layout is not a direct inventory of universally agreed linguistic units.
Formation of new lexemes
Derivation typically creates a new lexeme from an existing lexical base. HAPPY, UNHAPPY, and HAPPINESS are therefore normally treated as distinct lexemes, even though their meanings and forms are related.
Compounding also creates lexemes, as in TOOTHBRUSH. A lexeme can consequently contain several roots and have a meaning that is partly or largely predictable from its components.
A new lexeme may also arise without an overt change in form. The noun bottle and the verb bottle, meaning to put something into bottles, are commonly analyzed as distinct lexemes related by conversion. Their identical spelling does not erase their grammatical differences.
The boundary between inflection and derivation is not equally clear in every language or theoretical framework. Participles, verbal nouns, and some comparative formations are among the cases for which analyses differ. The general distinction remains useful: inflection realizes grammatical contrasts within a lexeme, whereas derivation establishes a relationship between lexemes.
Meaning and lexical identity
One lexeme may have several related senses, a phenomenon known as polysemy. The English noun mouth, for example, can refer to a bodily opening or to the opening of a river. Related senses do not automatically require separate lexemes.
By contrast, homonymy involves distinct lexical units that share a form. Bank meaning a financial institution and bank meaning the side of a river are ordinarily treated as separate lexemes.
The division between polysemy and homonymy is not always sharp. Historical origin, speakers’ judgments of semantic relatedness, grammatical behavior, and dictionary practice can support different classifications. Historical relatedness alone does not establish that contemporary speakers treat two uses as senses of a single lexeme.
Lexical identity also need not correspond to a single space-delimited word. Fixed expressions such as by and large can function as conventional vocabulary units. Terminology varies: some descriptions permit multiword lexemes, while others reserve lexeme for word-level units and use broader terms for fixed phrases and idioms. Neither the presence of spaces nor semantic unpredictability alone settles their status.
Cross-linguistic scope
The lexeme–word-form distinction is especially visible in languages with extensive inflection. A single lexeme in Latin, for example, may have many forms expressing contrasts of case, number, tense, or other grammatical categories.
In languages with relatively little inflection, the distinction may be less conspicuous, but questions of lexical identity remain important. In Chinese, a written character does not automatically correspond to a lexeme: many vocabulary items contain multiple characters, and the relationship among writing, morphemes, and words requires separate analysis.
The concept is not restricted to spoken or written language. Descriptions of sign languages also distinguish lexical units from their contextually or grammatically conditioned realizations, although the relevant morphological patterns need not resemble those of spoken languages.
Uses and analytical limits
In lexicography, lexemes help organize vocabulary into entries and distinguish inflected forms from separately listed derivatives. In corpus linguistics, grouping forms by lexeme makes it possible to study a vocabulary item across its grammatical realizations rather than counting each spelling independently.
Lemmatization assigns word tokens to conventional citation forms. This often approximates grouping by lexeme, but the two operations are not identical. A system that assigns every occurrence of bank the lemma bank has not necessarily distinguished its different lexical identities. Context, grammatical category, and semantic analysis may be needed.
Lexemes are analytical abstractions, not necessarily discrete objects stored unchanged in a speaker’s mind. Theories differ over the roles of stored words, morphological rules, constructions, and learned patterns. Likewise, counts of a language’s lexemes depend on decisions about compounds, technical vocabulary, regional variants, fixed expressions, and related senses. A lexeme count therefore requires explicit criteria rather than simply a count of distinct written forms.