Terminology for a group of words derived from a common stem?

Question

What is the linguistic term for a palette of words, which are derived / span from the same stem (excluding the so-called doublets like warranty vs guaranty...)?

Example:

Act (verb), react (antonym?), action (verb-noun), actor/actress (a creature doing the verb), active (adjective), etc...

PS I don't count explicitly the various inflections of a word in the span (e.g. (s/he/it) acts, (many) actors, ...) as part of what I call a palette and don't consider pairs of words like "see" and "sea", which sound similarly but don't have any etymological relation, as palettes.

Answer 1

If you're wanting a term which will refer to the set of words which are related through current derivational processes and have a current shared meaning, then these related words could be said to be part of a word family.

A word family is the base form of a word plus its inflected forms and derived forms made from affixes (Hirsh & Nation 1992, p. 692).

If you also want to include doublets, words in one language which share an etymological source but may not seem connected any more, such as, apparently, grammar and glamour, then I'm not sure if word family would be an appropriate term. All derived words, cognates and doublets are related and in a sense part of a family, but people might not think they would all be included when they think of a word family. To be safe I'd recommend specifying if you are including cognates or doublets.

Answer 2

It isn't clear what you mean by "derived from". For example "father" and "paternal" historically derive from the same root in proto-Indo European, but in English there is no derivational relationship between these words. Words that are related in that sense of "derive from" are referred to as "cognates" (e.g. "these words are all cognate(s)"). The term generally refers to related words across languages (so would include Sanskrit, German, Slavic and so on), but you can talk about the set of words in English that are cognates, thus father, paternal but not πατηρ.

In terms of synchronic notions of "derive from", where e.g. "digging" and "digs" are derived from the same stem, the term "paradigm" is used to cover the relationship, so the set of words morphologically derived from a given stem would be the paradigm of that stem. One can further subdivide paradigms into sub-types, like derivational vs. inflectional paradigms.

It looks like you're talking about derivational paradigms. However, it is dubious that there is a synchronic relation between your specific examples – there isn't a clear rule-governed relation between the specific forms. There are a number of nouns derived from verbs that add -ion, but the form and meaning of those words is too disparate to be subsumed under any rule (see for example rebel, rebellion; repel, repulsion; divide, division; unite, union; torque, torsion; cite, citation). We generally consider such sets of words to not be synchronically related. Which suggests that you are interested in historical relations, not synchronic morphological ones.

The concept of "cognate" in historical linguistics generally refers to the entire set of words having a common source, so the set of English words that are in this cognate set would include agent, agitate, agony, pedagogue, axiom and a number of other words. You could narrow down the set of words covered by the concept "cognate" by specifying for example "coming from Latin" or "coming from the Latin past participle". There is no single term that would cover exactly the words that you are interested in, at least based on your examples.

Answer 3

In the European linguistic tradition, the term word family (or morphological family) is the most common way to refer to a group of words having the same root (e.g. Booij 2007: 4, Haspelmath and Sims 2010: 17), cf. Bauer 2004, A glossary of morphology:

word-family: A word-family is a set of lexemes derived from the same base. Father, fatherly, fatherhood, fatherland, godfather are all members of the word-family based round father.

Dressler and Ladanyi 2000 argue there is a difference between a word-formation nest ("all synchronic and morphosemantically identifiable derivations from the same base") and a word family (both synchronic and diachronic derivations).

It is also the most usual term in language acquisition studies (see curiousdannii's post above with a quote from Paul Nation).

Less common terms: derivational family, lexeme family.