Numerals and other quantifiers may act like determiners in some circumstances the beeves these beeves thirty beeves but this is mainly due to the lack of a plural indefinite article. However, they can coexist with determiners, and determiners cannot coexist with each other the thirty beeves these thirty beeves *the these beeves On this basis, and allowing for the fuzziness of English parts of speech, I'd say that numerals behave more like adjectives than anything else, so they're not function words. I'm intrigued by your assertion that there is no category of preposition. How would you classify the words that are usually considered prepositions in English? ----Original message---- From : [log in to unmask] Date : 19/09/2017 - 18:06 (BST) To : [log in to unmask] Subject : Re: [CONLANG] Why doesn't English permit tl or dl as true consonant combinations? On 19 September 2017 at 15:21, PETER BLEACKLEY <[log in to unmask]> wrote: It seems that sychronically, the rule is one of restriction, rather than prescription. This goes without saying. ð may only occur as the initial consonant in function words where the second phoneme is a vowel. This is true. Function words include pronouns, demonstratives and conjunctions, but not numerals. This definition of function word is certainly not necessitated by the previous generalization. The previous generalization still holds even if numerals are function words. The definition of function word is of course open for debate, but I would very much want to argue for a definition that includes numerals for the units and decads. The question of whether _thirty_ and _thirteen_ are function words becomes relevant only if you want to posit a rule that says "word-initial onset th is voiced iff in a function word". But I would argue that such a rule would be mistaken, not only because it entails an untenable definition of function word but also because it fails to adequately capture the phonology of initial eth and also because it'd be a kind of rule that, in terms of the sort of interaction between phonology and syntax it entails, is unheard of anywhere else in English. I think they would also include prepositions, but the only preposition with an initial dental fricative that I can think of is through, which has a consonant as the second phoneme. Yes. (Some people consider _there, thither, thence, then, though_ to be prepositions, btw. My own view is that there is no category of preposition.) Diachronically, weaker stress on function words is probably the cause, although my use of the word "unstressed" was probably an overstatement. Yes, that was presumably the case at the first diachronic stage of the change; but at some point the triggering conditions must have ceased to be prosodic and instead become morphosyntactic. In PDE it would be analogous to the tapping/glottaling of initial t, which affects few words but _to_ and the medial t in _sometimes_. --And. http://specgram.com/CLXXVI.2/05.equality.announce.html ----Original message---- From : [log in to unmask] Date : 19/09/2017 - 14:31 (BST) To : [log in to unmask] Subject : Re: Why doesn't English permit tl or dl as true consonant combinations? On 19 September 2017 at 14:18, Eyal Minsky-Fenick <[log in to unmask]> wrote: > On Tue, Sep 19, 2017 at 8:40 AM, And Rosta <[log in to unmask]> wrote: > > > On 19 Sep 2017 09:04, "PETER BLEACKLEY" <[log in to unmask]> > > wrote: > > > > Diachronically, I predict the rule would be "Words that did not carry > > stress when þ and ð were allophones." In fact, sychronically, it may > simply > > be "unstressed words". > > > > > > Synchronically that is certainly false. In the unlikely event of the > > diachronic rule being correct, it would entail a definition of stress > > unrecognizable from the definition of stress in PDE. > > > > > > Numerals do carry stress, which is why "Thirty" is not affected. > > > > . > > "Thirty" is stressed like "thither" and "thicky" (demonstrative) and > > "thonder". Demonstrative "that" is stressed and does not habitually enjoy > > less accentual prominence than, say, "think". > > Wait, aren't those first two pronounced as /θɹti/ and /θɪðɹ/? > > /θɹti/ and /ðɪðɹ/ are the pronunciations I had in mind. But I discover, from checking OED, that there is indeed a US /θɪðɹ/, which then joins _thirty_ as an example of initial simplex /θ/ in closed-class words). --And.