Studying the book Understanding Morphology by Martin Haspelmath, I arrived at this fragment:

The importance of the latter part of the definition is seen in paradigms like insula. Although there are only seven different sequences of sounds in (2.3), we can still say that the paradigm of insula has ten word-forms, because ten different sets of grammatical functions are expressed (e.g. genitive singular and nominative plural are distinct, despite having the same form).

I thought a grammatical function was a relationship between constituents in a clause. So, my question is: why does the author gives genitive singular as an example? Is it supposed to be an inflection of the lexeme insula; a specific word-form?

  • Inflection is part of grammar. Morphology and syntax are both grammatical subjects.
    – jlawler
    Commented Jul 29, 2019 at 2:54
  • When I say in German "der Frau" (to the woman) or "der Mann" (the man), then I perceive these as different forms, perhaps because of different stress in different PoS, or because the form can be filled with different shapes depending on context. That has an ansser elsewhere, concluding that different ideas exist in the literature.
    – vectory
    Commented Jul 29, 2019 at 5:23
  • Interesting whether you (or Haspelmath) would count das Messer as one or two forms, since it can be either nominative or accusative, which count as two forms for masculine and feminine.
    – jlawler
    Commented Jul 29, 2019 at 15:17

3 Answers 3


Not having read the book, I find the terminology misleading in the same way you did. Nominative plural, genitive singular are commonly called grammatical features. For a constituent to have a specific grammatical function, it must have the proper grammatical features, e.g. for insulae to be the subject of the following sentence, we know it must be nominative plural and not genitive singular.

Insulae magnae sunt. The islands are large.

Or, the other way around, if insulae is to be interpreted as a postmodifier of incolae in the following example, we know it must be genitive singular.

incolae insulae
the inhabitants of the island

The first insulae could be replaced by urbes, the second by urbis, the nominative plural and genitive singular respectively of urbs "city".

I see no reason to collapse the distinction between grammatical features and grammatical functions. If nominative singular is a grammatical function, what do we call subject?


The terminology is just a bit unclear here. The real distinction is between different levels of the grammar.

On the phonological level and lower, the word insula has seven different forms (insula, -ae, , -am, -ās, -īs, -ārum, in no particular order). These are the "underlying forms" that phonologists are concerned about, which get processed through some sort of mental machinery (rules, constraints, whatever analysis you use) to produce the actual sounds.

On the syntactic level and higher, the word insula has fourteen different forms (nominative, genitive, dative, accusative, ablative, vocative, and locative; each of those can be singular or plural). The syntax doesn't care about the difference between Classical insulae and archaic insulās; all that matters is that it's the genitive singular form.

The morphological layer is right in the middle of these two groups, and it's what maps between abstractions like "insula genitive singular" and underlying forms like "insulae". When the author uses "grammatical function" here, what they mean is the abstract form of a word, the sort you see in glosses: something like insula-GEN.SG.

(P.S. The author considers only five cases, but there are technically seven; two of them were just in decline by Classical times, so they're sometimes ignored for simplicity. You can find grammars which list five, six, or seven, depending on how much they're willing to simplify.)


I read and used this book when I was a grad student.

I’m not sure how you could have missed their definition of “grammatical function” or “grammatical meaning” they give on the same page. The authors of the book - Martin Haspelmath and Andrea Sims - use these terms synonymously.

I don’t find their terminology misleading at all. Grammar could mean so many different things, depending on your theory of language (Universal Grammar, morphosyntax, etc.)

For instance, arguments such as Subject, Predicate etc. are known under different names, depending on your theory of language: grammatical relations, grammatical functions (in LFG), or syntactic functions etc. - see any intro syntax textbook.

  • Thanks for your input. I can’t find the definition of grammatical functions on that page. Can you point me to it?
    – F. Zer
    Commented Aug 2, 2019 at 15:10
  • 1
    Haspelmath and Sims 2010, pp. 15-16. Mea culpa - not exactly a definition there, but they make their point re: terminology very clear there.
    – Alex B.
    Commented Aug 2, 2019 at 21:12

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.