Configurational language where part of speech is marked?

[xBlackHeartx]

Yes, I know about the 'random conlang questions' thread. I didn't know about it for my previous two posts. And this one, I think may require a more in-depth discussion.

Its an idea I had. A flaw in configurational languages is that it can be hard to determine the part of speech of everything. The rules for determining what roles everything plays in the sentence are insanely complicated, and even then ambiguity issues still arise.

I thought of a way to solve this: mark everything for part of speech. If say the language had adjectives come after nouns, then the noun would always mark the beginning of the noun phrase. So if it was somehow marked as a noun (such as in Esperanto), then parsing would be easy.

Of course, I don't know of any natural languages that mark part of speech like Esperanto does.But many languages do mark part of speech on at least some things, they just don't use specialize affixes like Esperanto does. You're not likely to find an affix that just indicates something as a verb. Rather, a language may rely on tense endings. If only verbs can take tense endings, then a word having a tense ending would automatically mark it as a verb. If nouns have cases (such as in Latin), then a word having a case ending would automatically imply a noun. Yeah, there's Tok Pisin which has an adjective suffix, though its often omitted in everyday speech. Regardless, part of speech affixes are rare, since you can just double up with endings that words are required to have anyway. I mean, why have a suffix for 'verb' when the tense marker would mark it anyway since only verbs can have one? And yes, there are languages where other parts of speech can have tense. Japanese allows adjectives to have tense, though their endings are different, and not all adjectives can even take tense marking. Also, only verbs can take honorific endings, adjectives can not, even when they act as a predicate (if they want such a sentence to be honorific, they just tack 'desu' onto the end, its also common to do that in other situations where there's no actual honorific ending to use).

I'm not trying to create an auxlang here (though I tend to design my conlangs like they are anyway, partially because that's how I got into conlanging, and besides, I think it saves work to keep things simple). I just thought it was an interesting idea. Yeah, if you want nouns to be marked as nouns then you can just use articles or case endings. But this seems kind of a compromise. Nouns are explicitly marked as nouns, but you don't have to deal with case endings or articles. Though admittedly you could also just have your nouns have a marked singular. A plural ending automatically implies a noun (in most languages anyway, Korean verbs can take a plural ending, to mark a distributive aspect). This means that having a noun ending too would be redudant. So if the language had grammatical number, then there wouldn't likely be a noun suffix (asides maybe for word derivination) but instead just singular and plural markers.

[Dormouse559]

Could you please restate your post in three sentences or fewer? I'm having trouble teasing out what you want to discuss.

[Bubulus]

He's discussing strategies to unambiguously mark word classes, without having affixes that specifically mean NOUN, VERB, ADJECTIVE, ADVERB, like Esperanto -o, -i/-u, -a, -e.

As far as natural languages go, I don't personally see any advantage (or lack of a flaw) in non-configurational languages that mark word classes unambiguously. Classical Latin is quite non-configurational and has lots of curious syntactic ambiguities precisely because of it. It doesn't matter much that you can identify what word class a word belongs to when you have greater trouble identifying the components of constituents at all.

Filius Marci sororem vidit.
'Marcus's son saw his sister.' <- [filius Marci] sororem vidit
'The son saw Marcus's sister.' <- filius [Marci sororem] vidit

In Classical Latin, genitive nouns appear before a noun they modify about 50% of the time, and 50% of the time they appear after, so the sentence can really be interpreted both ways. As far as I know English doesn't have this ambiguity at all thanks to having a much stricter word order.

I don't see much of a problem with having a language with rather strict word order where word classes are easy to identify, especially if it's done through number/TAM marking or articles and the like. In fact, Classical Arabic could probably count as one of such. (Adjectives in Classical Arabic look just like nouns in apposition, but they're not hard to identify because of their syntax anyway.)

[xBlackHeartx]

Sigh. My main question is: is there any benefit to explicitly marking part of speech?

I've seen many configurational auxlangs be criticized for lacking this. Probably the most common is the lack of distinction between object and possessive pronouns. For example, in Lingua Franca Nova, 'mi parla' can either mean 'I talk' or 'my talking'. Note that in the case of this example, this problem has been fixed. 'My talking' would now be 'ma parla'. Though I do recall running into the same problem with Glosa, and a few others I can't think of right now.

Another example I can think of is Toki Pona. In Toki Pona, this problem comes up all the time. Particularly with 'li', which can either mean 'to be' or indicates the following as a verb. Thus 'li moku' can either mean 'to eat' or 'to be food'. Problems like this happen all the time, though Sonja Lang (I think that's her name?) does mention this problem, but just ignores it. She didn't really intend Toki Pona to be an auxlang anyway.

As for my original post, I kinda went on a tangent about natlangs that do mark part of speech don't just have affixes specifically devoted to it. Esperanto is literally the only language I know of that has affixes that just mean 'noun' and 'adjective' and 'infinitive verb' and such.

So Esperanto doesn't have case endings (aside for the accusative of course), but still has an ending there just to indicate that something is a noun. But is this really necessary? To my knowledge Chinese gets away without such things. They don't even have articles (to me, it seems like English relies heavily on articles to indicate the beginning of a noun phrase)! But is there actually a benefit to marking part of speech like Esperanto does?

Also note that I don't actually know Chinese. What little I do know comes from reading the LCK 2 and studying Kah, which to my knowledge uses a grammar that's just Chinese grammar, but heavily simplified.

[Bubulus]

Sigh. My main question is: is there any benefit to explicitly marking part of speech?

Sure, there is. It makes parsing easier. I don't think it's necessary though, considering the very large number of languages that happily use plenty of zero-derivation (English and Mandarin being prominent examples).

I've seen many configurational auxlangs be criticized for lacking this.

You make it sound like that's a problem. Auxlangs will always be criticized for having this or not having that, it's unavoidable. Five years ago I once saw a Japanese conlanger come to the ZBB and criticize Westerners' conlangs in general (auxlangs and artlangs!) for commonly having the singular-plural distinction in nouns, which he considered utterly unnecessary, and in fact, outright stupid. Yet if you made an auxlang without the singular-plural distinction (you know, following certain languages spoken by massive numbers of speakers, like Mandarin, Malay-Indonesian and Bengali), I'm sure there would be waves of conlangers, whose mother tongue is a Western language, criticizing you for it. Damned if you do, damned if you don't.

Probably the most common is the lack of distinction between object and possessive pronouns. For example, in Lingua Franca Nova, 'mi parla' can either mean 'I talk' or 'my talking'. Note that in the case of this example, this problem has been fixed. 'My talking' would now be 'ma parla'. Though I do recall running into the same problem with Glosa, and a few others I can't think of right now.

That sounds like a lack of distinction between subject and possessive pronouns. Mandarin doesn't always make it either (wo3 'I, me', wo3 de 'my', wo3 ma1ma 'my mother', wo3 ge1bo 'my arm', wo3 de lin2ju1 'my neighbour'), and in fact it also uses complete sentences as the subject of another sentence all the time:

我们丢失衣服让我感到很懊恼。
wo3men diu1shi1 yi1fu rang4 wo3 gan3dao4 hen3 ao4nao3
1PL lose clothes let 1SG feel very upset
(Somewhat literally: "us losing the clothes lets me feel very upset")
'I feel upset because we lost the clothes [that we had bought].'

"1PL lose clothes" is a complete sentence that can mean 'we lose the clothes', and yet it is used as the subject of rang4 'let'. So, more literally, this is "[we lose the clothes] lets me feel very angry". Distinguishing "I talk" from "my talking" is not really necessary, and might even be desirable because of the many Mandarin speakers who don't make this distinction either!