I just read that usually word formation processes occur before any inflectional processes. Are there any languages in the world where the opposite happens?

  • Depending on what you mean by before, there are the Mayan languages -- highly inflected, completely prefixal. So all inflections occur before anything else, let alone word formation.
    – jlawler
    Aug 19, 2023 at 22:03

3 Answers 3


Linguists usually treat inflection as a type of word-formation, hence "sees" is a word, "see" is a word, and "seeable" is a word. The dichotomy that you are probably thinking of is between "inflection" and "derivation". There was a large literature on the question of what characterizes "derivation" as opposed to "inflection", typically centered around "lexical / core meaning" vs. "grammatical function". Grammatical function typically is seen as those properties that must be marked on a word in order for the word to stand in a particular slot, for example "as subject" (relation to predicate), "modifying a feminine plural noun", and so on. Agreement, syntactic-frame, and temporal reference are canonical examples of inflection, and processes that mark those kinds of properties are often relegated to the status "form of a word", rather than "distinct word".

When you consider a broader range of languages than the major European languages, you will encounter processes that are hard to classify by the traditional criteria. For example, many languages have a "causative" verb form that allows a bare NP to serve as the object of the main verb, when that NP is the subject of the main verb, and someone else makes the NP do it – e.g. "I made the person cook" which in Shona is ndakabikisa munhu, literally "I-made-cook person" (bik is "cook"). This allows a bare NP to be integrated into a syntactic frame in novel ways, but it has a semantic function as well, and is very similar to the derivational prefix en- in English. The noun class prefixes in Shona similarly serve as indicators of number and are properties that trigger obligatory agreement (adjectives have to agree in noun class with the head noun, idem verbs with subjects), but the classes also reflect typical non-syntactic properties (diminutive, augmentative) and are also lexically-determined (bhasikoro "bicycle" is lexically assigned to cl. 9, badza is lexically assigned to cl. 5, you just have to learn it as part of the lexeme). There are regular prefixes on verbs meaning "to V early", ones used to plead with an interlocutor, or to express the idea of distance (long ago, highly unlikely, far away). These are not your typical European inflectional affixes, but they are also nothing like European derivational affixes.

The most useful dichotomy that can be made in this area is between lexicalized forms, where the function and meaning of a morpheme is not purely compositional, versus compositional. Syntax is a rule system that mediates between superficial form and meaning, meaning that by definition such relations are regular. There are problematic cases such as phrasal idioms (kick the bucket, take a shot at) and highly-lexicalized verb+partical constructions involving two words (look up, get across (a meaning), carry out (implement)), which indicate that the fact of being "a word" is not grammatically very significant.

No language with more than trivial morphology requires all lexically-restricted affixes to appear "outside" all syntactically-governed affixes. Many languages allow some syntactically-governed affixes to appear "inside" lexically-restricted affixes.

  • 2
    Inflection is not word formation, it is at best "word-form formation". I have never seen this kind of broad meaning for the term "word formation" that you mention, where do you have that from?
    – Alazon
    Aug 18, 2023 at 16:16
  • @Alazon However boundary between derivation and inflection can be fluent. E.g. in modern Japanese, Perfective suffix -ta is only considered to be inflectional suffix, but Causative suffix -(s)ase- and Passive/Potential/... suffix -(r)are- can be considered to be inflectional or derivational suffixes. Morphologically they are more derivational, but since meanings of words with these suffixes normally can be predicted, they do not usually have separate entries in dictionaries. (E.g. yomu "to read", yomareru "to be read", yomaseru "to make (somebody) read")
    – Arfrever
    Aug 18, 2023 at 16:58
  • Also treatment of nouns, adjectives (adjectival participles), adverbs (adverbial participles) derived from verbs, nouns derived from adjectives etc. often varies between languages. In English, sadsadness is considered derivation, but in Japanese kanasii (悲しい)kanasisa (悲しさ) is considered part of inflection of this adjective.
    – Arfrever
    Aug 18, 2023 at 17:06

That's not so much a matter of language or language type, but a matter of individual examples. Your book says "usually" because the generalisation is very strong, it is almost taken as a definition of inflection that it is peripheral. If there were languages that systematically violate the generalisation in all of their grammar, it would not have been formulated this way (one would hope :).

However, there may be problems with the word "before" which you used. An apparent systematic counterexample are particle verbs. They have a structure in which inflection appears to be closer to the verb stem than the particle. That's true of English, it's even more impressive in German, in which the particle looks even more like word formation ("zu" and "ge-" are markers for different types of infinitive, among other uses):

  • an + ziehen (= put on)
  • an + zu ziehen (to put on)
  • an + ge-zogen ("participle")

However, particle verbs could be viewed as idioms consisting of two words. On such a view, it is not a problem that the inflection comes "first", i.e. applies to the simple verb and not to the whole particle verb. If particle verbs are viewed as word-formation, it would be a problem (and this is more pressing in German because particle verbs can feed further word formation very easily). But it is not clear that "word-formation" is a unified notion, just as "word" is not a unified notion, it is rather ambiguous.

One could flatly view the German data as counterexamples - or one could react by refining the grammatical description and the exact scope of the ordering rule.


Japanese has compound verbs consisting of multiple verbs joined together.

Before derivation of such compound verbs, each non-last verb must be inflected to its Continuative/連用形 form.

During inflection of such compound verbs, only last constituent verb is inflected (e.g. gains some suffixes).

In compound verbs, first constituent verb often specifies general area of meaning, and subsequent constituent verbs modify meaning, sometimes acting similarly to prefixes (etymologically coming from prepositions) of verbs in Latin, Greek, Slavic languages.


yomu (読む) ("to read"), Continuative/連用形 form: yomi (読み)

yomu (読む) + kakeru (掛ける) ("to hang up, to let hang, to suspend, ..., to be partway doing ..., to begin (but not complete) ...")yomikakeru (読み掛ける) ("to begin to read, to leaf through a book")

yomu (読む) + kiru (切る) ("to cut")yomikiru (読み切る) ("to finish reading, to read through")

yomu (読む) + toku (解く) ("to solve, to work out, to answer, ...")yomitoku (読み解く) ("to read and understand, to interpret (e.g. a difficult text), to read deeply, to read carefully, to analyze, to decipher, to decode")

yomu (読む) + tobasu (飛ばす) ("to let fly, to make fly, to send flying, to skip over, to leave out, to omit, to drop, ...")yomitobasu (読み飛ばす) ("to skip over (while reading), to skim")

yomu (読む) + nagasu (流す) ("to let flow, ...")yominagasu (読み流す) ("to read smoothly, to skim")

Inflection of these verbs to Perfective form (usually past tense meaning) (which is formed with suffix -ta, sometimes voiced to -da, and sometimes causing sound change in preceding syllable):

yomu (読む) → yonda (読んだ)

yomikakeru (読み掛ける) → yomikaketa (読み掛けた)

yomikiru (読み切る) → yomikitta (読み切った)

yomitoku (読み解く) → yomitoita (読み解いた)

yomitobasu (読み飛ばす) → yomitobasita (読み飛ばした)

yominagasu (読み流す) → yominagasita (読み流した)

Continuative/連用形 form has also other functions than derivation of compound verbs, e.g. standalone non-finite verb form, and base for some inflectional suffixes.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.