Do all words have a part of speech? The closest counterexample I can think of is yes. The dictionary says its supposed to be an adverb but it doesn't really strike me as something that modifies a verb. Are there any words in common use that simply don't have a part of speech? Or, alternatively, are there any parts of speech that they just don't teach in school because they're too abstract to the children learning their parts of speech?

  • 6
    This isn't my area, but I sort of always assumed that if a word didn't have a part-of-speech the linguists would create a new part-of-speech category for it.
    – MrHen
    Jul 9, 2011 at 19:45
  • 1
    This question is interesting, but... Isn't it more fit to be on the Linguistics SE?
    – Alenanno
    Jul 10, 2011 at 9:20
  • 1
    I'd call yes an interjection. // Adverbs need not modify verbs: they regularly modify adjectives and other adverbs as well. A very large house. A surprisingly accurate prediction. They went very fast and came too soon. Jul 16, 2011 at 21:15

3 Answers 3


It is intended that all words have a part of speech.

As for your second question - "are there parts of speech they don't just teach in school?" Traditionally, The eight parts of speech are: nouns, verbs, pronouns, adjectives, adverbs, prepositions, conjunctions, and interjections. Students learn all of these.

However, depending on what we're doing, we can use a lot more different types of speech. For example, for my natural language processing project, I used 44 different types of speech (including some like "numbers" and "periods")

  • Thanks for the fully explained answer. +1 for the 8 parts.
    – Jamie
    Jul 10, 2011 at 9:50

Strictly speaking, words don't have parts of speech or indeed any properties at all.

Look at this from the point of view of Luciene Teniere's 'valency grammar', in which a sentence is a sequence of slots with a function, and constructed by dropping appropriate fillers (words or phrases) into the slots. Very briefly, "cats catch mice" has three slots - subject, filled by a doer (cat). Verb, filled by what is done (catch), and object (mice), that is filled by the victim of what is done. If the cats are big hairy salivating ones, then "big hairy salivating cats" would fill the object slot.

The point here is that the 'noun' characteristics of 'cats' do not come from any quality of the word itself, but from the slot it fills in the sentence and how it functions there. This can be seen when 'wrong' words are dropped into slots. A gerund? An -ing verb dropped into an object slot. Elsewhere on this site there is a question about adjectives being used as nouns, with 'The successful are those who strive' as an example. I pointed out that 'the sucessful' is a nominal adjective rather than a noun, but that invites the question "what is a nominal adjective?" It is an adjective being dropped into the object slot of a sentence where we would expect to see a noun.

By putting these words in object slots they have to function as a name of something - become a noun - so we have the same orthographic word functioning as different parts of speech depending which sentence slot it is put in.

  • That may be philosophically interesting, but pretty much always 'cat' is noun-like (except when you're catting around). It's pretty useful to just say that 'cat' has the property 'noun'. 'cat' doesn't serve any purpose as a preposition (yet!) 'The man cat the roof' could eventually come to mean the man is up on the roof (like French 'chez Henri' means 'at Henri's place' coming from Latin 'casa de Henrius' = 'house of Henry')
    – Mitch
    Mar 1, 2012 at 16:35
  • @Mitch: Well, there is catting, as you said, and a cat bell is not a cat, but a cat-related bell, where cat serves more or less as an adjective (as most nouns can, but of course we all know that). So while the label "noun" is by all means useful and practical, there are always qualifications that one can make. As Fish says, parts of speech are convenient labels we assign to (instances of) a word, and nothing more. This applies to any linguistic term, of course, so you could say it is a commonplace; but it seems a useful bit of information for the question asker. Dec 17, 2012 at 1:40
  • @Cerberus: Depends on what 'is' is. A word can be one (or more) parts of speech or maybe it just acts like another part of speech while remaining a canonical one, or maybe a mix.
    – Mitch
    Dec 17, 2012 at 13:34
  • @Mitch: Yes, exactly: a word "is" nothing but what we make of it. Dec 17, 2012 at 16:38

(1) Classifications of POSs vary enormously, but no grammarian would see the traditional 8-class model as satisfactory. But many in-house approaches are used as found most appropriate by institutions and departments (and new prescriptivists seem not to appreciate that there may be better alternatives to their own pet models. Research is an ongoing process.) You can imagine how hard it is to find a single workable simplified model for schools, one that will not soon give rise to major problems.

(2) Some words seem to fall between two (or more) prototypical classes. A word's POS (part of speech) is determined by some mixture (and the weightings can be more or less arbitrary) of (a) their formal properties (eg does it pluralise by the addition of an s?), (b) their distribution (does it always fit in a slot like 'This is ___ dog / picture / mess ...?), (c) further grammatical considerations. Even so, just to take one well-known area which affords a prime example, many ing-forms seem to be somewhere on a noun ... verb continuum (some workers use this 'gradience' model, for instance Quirk et al in ACGEL, while some are more inclined to lump, or even put a word into what is considered the nearest prototypical class) while it can be argued that 'galore' say is both quantifier and adjective at the same time. Splitting (more POSs) / lumping (Fewer, more diversely populated POSs) is a judgement call; Aarts looks more deeply at this.

(3) I leave most idioms as indivisible units (lexemes: after Crystal). I'd class 'Yes' as 'sentence substitute', as 'interjection' and 'exclamation' surely imply / strongly connote an abrupt interruption: not always the case. This conflates single orthographic words (Please!' / 'Yes.' / 'Yes!' / 'Never!' / 'Wow!' / 'Oops./! / ...) with longer strings ('Yes, please.' / 'On the table.' / 'When I'm ready.' ....) Far from ideal, as I can't really class them as idioms and invoke the massive endorsement of Crystal. I'm still working on this.

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.