A polysynthetic language is a language where words are made with lexical morphemes (substantive, verb, adjective, etc) as if parts of sentences were bound together to constitute one word, which can sometimes be very long. Those “words” will be translated by several words or even by a complete sentence for less synthetic languages such as English.

For instance, in Iñupiak-inuktitut, language of the Eskimo-Aleut family in North America:

Tusaatsiarunnanngittualuujunga means “I can’t hear very well”.

This word could be broken down as follows: the root Tusaa– (“to hear”) followed by 5 suffixes:

tsiag– (“well”), –junnag– (“to be able to”), –nngit– (negative form), –tualuu– (“a lot”), –junga (marker of the first person and present tense).

Polysynthetic languages are usually agglutinative. Note that this definition of polysynthetic languages is also controversial. If the concept of polysynthetic language (and agglutinative) is operational for languages with very long “sentence-words”, as for Iñupiak, it is often difficult to tell how polysynthetic a language is if there are no written translations.

Nowadays, this definition is often used to describe any language where words are made out of a lexical/verbal root, when syntactic functions are only marked by affixes. In that case, we could describe Basque as a polysynthetic language, although there are no “sentence-words” as long as our example.

Note: the concepts of agglutinative and polysynthetic languages are controversial and their definitions are sometimes considered as non-operational.