zompist bboard

For example, in IE languages you tend to do something like:

[noun phrase], [noun phrase], [...] AND [noun phrase]

There's also the enclitic -que used in Latin, as in:

Senatus Populusque Romanus

What other options are used? For example, I imagine there's probably some language where the conjunction is required in between each word. But I'm interested in other options: are there options where the conjunction agrees with the noun phrases it joins? Are there languages where the nouns have to be inflected in a certain way to be conjoined?

In Malayalam, usually, you add a suffix to each noun phrase that's conjoined. So it's like [noun phrase]-[ʊm] [noun phrase]-[ʊm] [noun phrase]-[ʊm]...

Vijay wrote:In Malayalam, usually, you add a suffix to each noun phrase that's conjoined. So it's like [noun phrase]-[ʊm] [noun phrase]-[ʊm] [noun phrase]-[ʊm]...

Interesting. How would you handle nesting of conjunction in Malayalam? For instance: "ketchup and mustard, salt and pepper, and mayo and relish"

I don't think you would.

EDIT: Actually, there is one trick you can use for that: "ketchup-[ʊm] mustard-[ʊm], salt-[ʊm] pepper-[ʊm], then mayo-[ʊm] relish-[ʊm]."

Verbal/clausal conjunction and nominal conjunction may take different forms. Verbal is generally the same as clausal, but sometimes different, either with a three-way contrast or clausal versus verbal/nominal. It's fairly common to have an explicit nominal conjunction but mark verbal conjunction purely by juxtaposition.

Nominal conjunction is often identical to the adposition "with."

The line between "and" and "or," or further differentiation between the two, varies greatly between languages (page 30 w/source).

Someone mentioned on here a language where nominal conjugation inflects like a transitive verb, the first noun being the "verb's" subject and the second the object, but I don't remember what it was.

Some languages seem to lack a word for 'and', such as Middle Egyptian, and just adjoin the words. I've seen this in Thai for closely linked words, like 'father and mother'. In Sanskrit, one can form dvanda compounds, where all but the last word lose their inflections.

The word for 'and' can affect the following word, e.g. Welsh a(c) is supposed to take the aspirate mutation on the following word, so 'pen and paper' is pen a phapur. The use of a preposition meaning 'with' is widespread amongst languages, and that can have all the effects that a preposition normally does.

vokzhen wrote:Someone mentioned on here a language where nominal conjugation inflects like a transitive verb, the first noun being the "verb's" subject and the second the object, but I don't remember what it was.

I was just about to post this, it's Walman.

Might be worth pointing out that in french, CONJ NP CONJ NP has a special meaning:

et Jean-Pierre et son frère = both Jean-Pierre and his brother
ou bon ou mauvais = either good or bad

Logically, one would expect that CONJ NP NP would be attested somewhere, but I can't find any examples by quick Googling.

zompist wrote:Might be worth pointing out that in french, CONJ NP CONJ NP has a special meaning:

et Jean-Pierre et son frère = both Jean-Pierre and his brother
ou bon ou mauvais = either good or bad

Latin and Russian have that, too.

There's and and there's and. Not all conjunctions are equal and the nuances can be quite tricky to point down.

Icelandic has no independent conjunction akin to English "yet", conflating it with en which primarily means "but". But Icelandic en can be used to form semi-dependent clarification clauses, usually translated into relative clauses in English or replaced by a dash:

Nýji starfsmaðurinn kom í gær en hann hafði áður unnið hjá PepsiCo
"The new employee came yesterday – he had previously worked for PepsiCo"

So that's weird.

Icelandic also has the conjunction enda which has no English equivalent, meaning something like "as", "granted that", "which as you know" or "which makes sense as" something like that. It's especially weird in that it demands V1 word order, moving the verb in front of the subject:

Nýji starfsmaðurinn kom í gær enda var samninginum hans hjá PepsiCo lokið
"The new employee came yesterday which makes sense as his contract was up at PepsiCo"

So that's also weird.

Latin had like four-five ands: -que, et, atque, ac, nec non. The last one's more like "and besides" and atque is sometimes more appropriately translated as "as well as" or "and moreover" but not always. It also had two ors, sed and aut. But it has no "yet" or "enda".

Greek has those particles like oun and de which are sort of like conjunctions but a little bit like discourse particles, too. And they appear as the second word in a sentence, breaking up even a noun from its article.

There's a lot that can be done with conjunctions…

vec wrote:Latin ... also had two ors, sed and aut.

Latin had two ors, vel and aut.

A vel B allows the possibility of A and B; A aut B doesn't. This distinction eventually vanished. A vel B could also be said A Bve; I'll let you decided if that's a third 'or'.

vec wrote:Icelandic has no independent conjunction akin to English "yet", conflating it with en which primarily means "but". But Icelandic en can be used to form semi-dependent clarification clauses, usually translated into relative clauses in English or replaced by a dash:

Nýji starfsmaðurinn kom í gær en hann hafði áður unnið hjá PepsiCo
"The new employee came yesterday – he had previously worked for PepsiCo"

So that's weird.

Not really; it works like the Latin words for 'but', and is similar to the use of English but in mathematical proofs, which I have always assumed to be a calque on Latin.

Note also that Ancient Greek δέ can mean "and" or "but". This actually makes sense—the meaning of the sentence is often enough to tell you whether it is consistent with, or contrary to, the speaker's expectations.

Haspelmath has a thing or two to say about conjunction cross-linguistically. Note that conjunction (with conjunctions and conjuncts) in his terminology specifically refers to conjunctive coordination (A and B). He uses coordination (with coordinators and coordinands) as the more general term, which also includes disjunctive coordination, or disjunction, (A or B) among other subtypes.
http://email.eva.mpg.de/~haspelmt/coord.pdf

Haspelmath was also the editor "Coordinating Constructions", which happens to be volume 58 the John Benjamins Publishing Company series "Typological Studies in Language".

Regarding the order of coordinands, there are nine logically possible orders for binary coordination: A B, A co-B, A-co B, A B-co, co-A B, co-A B, co-A co-B, A-co B-co, A-co co-B, co-A B-co.

All orders are apparently attested in natlangs except co-A B. For the type A-co co-B, Haspelmath gives an example from Homeric Greek:

Atreídes te kaì Akhilleús
Atreus’s.son and and Achilles
‘Atreus’s son and Achilles’

For the type co-A B-co, he gives a Latin example:

et singulis universis-que
‘both for individuals and for all together’

However, at least the Latin example is what Haspelmath calls contrastive coordination, ie "both A and B", "either A or B" (compare the French examples above). I'm not sure if there are languages that uses the co-A B-co for more neutral coordination. I'm also not sure how the Homeric example should be interpreted.

It is also interesting to note that many languages distinguish natural conjunction and accidental conjunction. In natural conjunction, the conjuncts "habitually go together and can be said to form some conventionalized whole or ‘conceptual unit", e.g. ‘mother and father’, ‘husband and wife’, ‘boys and girls’, ‘bow and arrows’, ‘needle and thread’, ‘house and garden’ (his examples). I don't know if there are any language that uses different overt coordinators for the two, but natural conjunction may involve lack of overt marking (Thai, as mentioned above) or lack of intonation breaks. They may even form compounds (Sanskrit, as mentioned above).

In English, natural conjunction requires only one article (the house and garden) while accidental conjunction requires two (*the house and stamp collection; examples again from Haspelmath).

zompist wrote:Logically, one would expect that CONJ NP NP would be attested somewhere, but I can't find any examples by quick Googling.

According to Haspelmath, that's actually the one order that's unattested.