NSM Natural Semantic Metalanguage

Natural semantic metalanguage (NSM) is a linguistic theory claiming that:

• all lexical items (lexemes) in any human language are made up of a small number of simple concepts (‘semantic primes’);
• the semantic primes cannot be broken down into more basic concepts;
• the semantic primes are universal: the same semantic primes underlie every language. Also, for each prime, every human language has a lexeme corresponding to that prime. (That lexeme is known as the exponent of the prime.)

NSM was developed by Anna Wierzbicka in the early 1970s at Warsaw University and then at the Australian National University, and Cliff Goddard at Australia’s Griffith University.

The semantic primes

Goddard (2011) lists 63 semantic primes hypothesized at that time. Table 1 lists them, using their exponent in English. In line with Goddard, the primes are gathered in various categories. As far as I can tell, the categories are used only for convenience and have no significance.

Table 1. NSM semantic primes (English exponents)

 Notes on table 1

1. The primes are given in all capitals, presumably to distinguish them from other lexemes.
2. The list of primes is still a working hypothesis and may change as work progresses. The resource base of publications using the Natural Semantic Metalanguage approach lists several items once analysed as semantic primes but no longer considered to have that status. They include HAVE, IMAGINE, UNDER, ON, WORLD.
3. Exponents of some semantic primes have different forms (‘allolexes’) in different contexts. Table 1 links allolexes of the same sematic prime with the symbol ~. In Table 1, 7 of the primes listed have more than one allolex. Their English exponents are: SOMETHING~THING; OTHER~ELSE; MUCH~MANY; BE (SOMEONE~SOMETHING); WHEN~TIME; WHERE~PLACE; LIKE~WAY.
4. Although most semantic primes have a simple form, made up of a single morpheme, this is not always the case. In table 1, SOMETHING, THE SAME, A LONG TIME, A SHORT TIME, FOR SOME TIME all have a complex form.
5. Semantic primes are not necessarily a complete word: in some languages, a semantic prime can be an affix.
6. Often, exponents of a semantic prime have additional meanings beside the semantic prime itself. For example, in some languages, the exponent of BEFORE can also have meanings such as ‘first’, ‘go ahead’ or ‘front’.  
7. Because the primes include NOT, A LONG TIME, THE SAME, GOOD, LIVE, FAR, it isn’t clear to me why A SHORT TIME, OTHER~ELSE, BAD, DIE, NEAR are needed.

Explications

NSM develops ‘explications’ of lexemes. These descriptions of semantic representations consist solely of semantic primes. Here is an explication presented in Goddard (2011) for the English phrase X was pleased at this time.

Semantic templates

Often, NSM explications for words of a similar semantic type have a common structure (semantic template).

For example, Goodard (2011) mentions the following possible semantic template for English emotion adjectives with the verb be:

In this template:

• section a describes the cognitive scenario.
• section b describes how the person feels as a result.
• section c characterises the feeling as typical of how people have when they have such thoughts.

Semantic molecules

In principle, all lexemes in a language can be explicated using only semantic primes. However, many NSM explications are very long.

To streamline the explications, NSM uses semantic molecules. NSM semantic molecules are explicated using semantic primes and the molecules are reused in explications of other lexemes. For example, one convenient semantic molecule is of ‘animal’. That molecule can be used in the explication of any particular type of animal (eg a cat, dog, horse or lion), without having to re-use the underlying explication of ‘animal’ in explicating words for types of animal.

Semantic molecules can be used in the explication of other semantic molecules, but ultimately each lexeme is explicated using only semantic primes, either directly or indirectly (through molecules).

Psychological plausibility

Proponents of NSM claim that its use of semantic primes corresponding to simple every day words makes it a plausible model of how children come to learn the meanings of words.

NSM beyond semantics

Apart from analysing the meaning of words and other lexemes, researchers have also used NSM to investigate, for example:

• syntactic topics, such as analysing the meaning of the dative case in Polish.  
• how meanings of verbs vary between when they are used transitively and when they are used intransitively.
• how behaviours vary across cultures, using what is claimed to be a universal and neutral metalanguage, rather than terms tied to a particular culture (such as the culture of anglophone countries).

Because the semantic primes are (claimed to be) universal, translators might find NSM helpful as a tool for investigating similarities and differences between languages.   

Other languages

Table 2 lists the primes, using their exponents in 4 languages: English, Japanese, Polish and Russian.

Table 2. NSM Semantic Primes: exponents in English, Japanese, Polish and Russian

Sources: Japanese and Russian: Goddard (2010), appendix A. Polish: Wierzbicka (2009)

Notes on table 2

1. Polish (a West Slavonic language) and Russian (East Slavonic) are closely related to each other. Not surprisingly, many of their exponents are similar.
2. Other allolexes of Japanese NANIKA (‘something’) are MONO and KOTO , but I am not sure whether they really have the same meaning.
3. I’m not sure whether Russian ROD (type, kind) and VID (aspect) can express the same meaning.
4. Japanese HITO / FUTA are from the native portion of the vocabulary and ICHI/Ni are Sino-Japanese. https://languagemiscellany.com/2026/02/numbers-in-japanese-and-chinese/
5. Several of the Russian exponents are separate forms for adjectives (XOROŠIJ, PLOXOJ) versus adverbs ( XOROSHO, PLOXO).  
6. Polish DUŻY (BIG) is an adjective and Polish
7. For WANT, Japanese HOSHI is an adjective. Other allolexes are -TAI (a suffix) and NOZOMU (a verb).
8. Another allolex of Japanese DOKO (‘when’) is TOKORO. Another allolex of Japanese OKORU is OKIRU.
9. Russian GOVORIT’(imperfective) and SKAZAT’ (perfective) are usually regarded as suppletive members of an aspectual pair, meaning ‘talk / say’). https://languagemiscellany.com/2026/04/suppletive-aspectual-pairs-in-russian/
10. For Polish, other allolexes are NIEKTÓRE (NIEKTÓRZY) and WSZYSTKO (WSZYSCY).
11. For Russian , another allolex of BOL’SCHE is EŠČE

Sources

Semantic Analysis: a practical introduction, Cliff Goddard (2^nd edition, 2011)

The Natural Semantic Metalanguage approach, Cliff Goddard, chapter 18 in The Oxford Handbook of Linguistic Analysis, edited by Bernd Heine and Heiko Narrog (2010)

Natural Semantic Metalanguage and Context, Anna Gladkova, chapter 10 in The Cambridge Handbook of Language in Context, edited by Jesús Romero-Trillo (2023)

Case in NSM: a reanalysis of the Polish dative, Anna Wierzbicka, chapter 10 in The Oxford Handbook of Case, edited by Andrej Malchukov and Andrew Spencer (2009)

A resource base of publications using NSM https://nsm-approach.net/