Fluid Noun Class Categories

[Mednij]

I've had the idea to have noun classes with the possibility of one word belonging to a different noun class depending on which quality is stressed. Then it turns out that the Niger-Congo languages beat me to it. For example, Fula forms diminutives and augmentatives by changing the noun class.

I decided to search for languages which have some "open" noun classes (where a noun can be a member of several different classes, unlike in many European languages with static membership), but could find little on the subject. The only such categories I could find are diminutives and augmentatives.

Do any of you know languages with "open" classes that assign other connotations?

[linguoboy]

I've had the idea to have noun classes with the possibility of one word belonging to a different noun class depending on which quality is stressed. Then it turns out that the Niger-Congo languages beat me to it. For example, Fula forms diminutives and augmentatives by changing the noun class.

I decided to search for languages which have some "open" noun classes (where a noun can be a member of several different classes, unlike in many European languages with static membership), but could find little on the subject. The only such categories I could find are diminutives and augmentatives.

Do any of you know languages with "open" classes that assign other connotations?

English. The use of she to refer to inanimate objects and of it to dehumanise people are both examples of this.

There are similar phenomena in other European languages. Spanish, for instance, makes a distinction between human animate objects and non-human objects of all sorts. But some speakers extend this class to include certain animals, generally those commonly kept as pets, e.g. Le vi a un perro (vs Lo vi un perro). And the pattern has apparently spread into Basque as well.

You can see something parallel in Swahili, where names of certain animals (generally class 9/10) take human animate agreement (class 1/2).

[Thry]

Though you can and should pronominalize direct objects with a as lo/la even though they have personal a beause they're still direct. Trust me, my D1 has this kind of leísmo. "Vi al perro; te juro que lo vi cruzar".

I exceptionally use le with *male* people (as a lo, never as a la), but I think with animals it's stretching it [in my dialect, ofc, then you have people who use this "le" even for potatoes, I've heard it].

[merijn]

I think there are a group of related phenomena, that are a bit mixed up in this thread.

1. A noun has can appear in multiple noun classes, and it has (if applicable) the noun class morphology of the noun class it appears in:
1a. A noun appears in multiple noun classes, and the meaning of a certain noun is completely predictable if you know the meaning of the stem, and the noun class. In the Niger-Congo language that I know best, Zulu, there are very few of those. The only group of nouns that I can think of is names of peoples: umXhosa (noun class 1) means Xhosa person, and isiXhosa (noun class 7) means Xhosa language or customs. I think that this actually is more common in Romance, where my guess is that something similar happens with nationalities, and I *think* that there are a group of nouns that differ in feminine or masculine gender, depending on whether they refer to a male or female person.
1b. A noun class appears in multiple noun classes, but the meaning is not quite predicable. In Zulu there are a lot of these. -hlobo for instance can be used with a variety of noun classes, all of them have to do with ‘being related to ’ but there is no way to know that noun class 1 (or 3, that is not clear to me) umhlobo means ‘acquaintance’ , and noun class 9 inhlobo ‘variety’.
1c. There is a noun class with a specialized meaning, and a stem of another noun class can be employed to derive a noun with that specialized meaning. This is essentially noun class functioning as derivational morphology. For instance, Zulu noun class 14 is mostly used for abstract nouns, derived from adjectives and other nouns. The noun ubu-ngane ‘friendship’ is derived from ‘um-mgane’ (noun class 1) ‘friend’. Other Bantu languages have specialized augmentative and diminutive noun classes.

2 A certain type of agreement is sensible to not only noun class but also to semantic information. This is what happens in Swahili with animate nouns from noun classes other than noun class 1 triggering noun class 1 morphology. Many Bantu languages have the same semantically conditioned noun class 1 agreement with animate (or human) nouns from other noun classes, but often this only happens with verbal agreement and not say with adjectival agreement. This is not radically different from German using the feminine pronoun sie for nouns that refer to female persons, even if the noun is not feminine, like Mädchen (but note that Mädchen is accompanied with the neuter article, and triggers neutral adjectival agreement), or British English use of plural subject agreement with singular nouns that refer to many people (‘the police have’). Sometimes there is no sharp line between the the case where you use one form or another, for instance, speakers may vary to what extent they use the animate agreement or the inanimate agreement with pets, depending on how much they see them as persons. Another thing is that people often play with this, one use that is particularly common is to use inanimate agreement with people to degrade them. This is what Linguoboy was talking about.

[Thry]

I've had the idea to have noun classes with the possibility of one word belonging to a different noun class depending on which quality is stressed. Then it turns out that the Niger-Congo languages beat me to it. For example, Fula forms diminutives and augmentatives by changing the noun class.

I decided to search for languages which have some "open" noun classes (where a noun can be a member of several different classes, unlike in many European languages with static membership), but could find little on the subject. The only such categories I could find are diminutives and augmentatives.

Do any of you know languages with "open" classes that assign other connotations?

mmm... I don't know if this is related but some words do have several genders: Spanish mar can be masculine (common noun, el mar) or feminine (poetic, la mar), with the exact same meaning. Other nouns that change gender also change meaning.

[Mednij]

I think there are a group of related phenomena, that are a bit mixed up in this thread.

1. A noun has can appear in multiple noun classes, and it has (if applicable) the noun class morphology of the noun class it appears in:
1a. A noun appears in multiple noun classes, and the meaning of a certain noun is completely predictable if you know the meaning of the stem, and the noun class. In the Niger-Congo language that I know best, Zulu, there are very few of those. The only group of nouns that I can think of is names of peoples: umXhosa (noun class 1) means Xhosa person, and isiXhosa (noun class 7) means Xhosa language or customs. I think that this actually is more common in Romance, where my guess is that something similar happens with nationalities, and I *think* that there are a group of nouns that differ in feminine or masculine gender, depending on whether they refer to a male or female person.
1b. A noun class appears in multiple noun classes, but the meaning is not quite predicable. In Zulu there are a lot of these. -hlobo for instance can be used with a variety of noun classes, all of them have to do with ‘being related to ’ but there is no way to know that noun class 1 (or 3, that is not clear to me) umhlobo means ‘acquaintance’ , and noun class 9 inhlobo ‘variety’.
1c. There is a noun class with a specialized meaning, and a stem of another noun class can be employed to derive a noun with that specialized meaning. This is essentially noun class functioning as derivational morphology. For instance, Zulu noun class 14 is mostly used for abstract nouns, derived from adjectives and other nouns. The noun ubu-ngane ‘friendship’ is derived from ‘um-mgane’ (noun class 1) ‘friend’. Other Bantu languages have specialized augmentative and diminutive noun classes.

2 A certain type of agreement is sensible to not only noun class but also to semantic information. This is what happens in Swahili with animate nouns from noun classes other than noun class 1 triggering noun class 1 morphology. Many Bantu languages have the same semantically conditioned noun class 1 agreement with animate (or human) nouns from other noun classes, but often this only happens with verbal agreement and not say with adjectival agreement. This is not radically different from German using the feminine pronoun sie for nouns that refer to female persons, even if the noun is not feminine, like Mädchen (but note that Mädchen is accompanied with the neuter article, and triggers neutral adjectival agreement), or British English use of plural subject agreement with singular nouns that refer to many people (‘the police have’). Sometimes there is no sharp line between the the case where you use one form or another, for instance, speakers may vary to what extent they use the animate agreement or the inanimate agreement with pets, depending on how much they see them as persons. Another thing is that people often play with this, one use that is particularly common is to use inanimate agreement with people to degrade them. This is what Linguoboy was talking about.

What I was trying to ask about was probably 1c or 2. To illustrate my point, I'll present a scenario. Imagine that we have a noun "cat" which is in the "animal" noun class. If we are talking about two cats, we can place one in a "diminutive animal" class and the other in the "augmentative animal" class to distinguish them. In neither of these cases does a separate lexical meaning develop (like kitten for the diminutive, or adult cat for the augmentative). If the small cat then leaves the big cat and goes to talk to a human, the size distinction isn't needed anymore and it now belongs to the neutral animal class. If it then goes on to talk to a mouse, the originally diminutive cat is now in the augmentative class and the mouse in the diminutive. Even when disambiguation like this isn't needed, a cat can be introduced as belonging to either the augmentative or diminutive noun class and later be referred to by the neutral class because the point was established. Thus, depending on which of a few basic quality distinctions is stressed (if any) (size or animacy, for example), a noun can belong to a neutral corresponding class or one that reflects one of these qualities.

My question was the following: Are there natlangs that can do such a thing besides with augmentative and diminutive properties (as well as with animacy and inanimacy as demonstrated by Linguoboy)?

[Yng]

In what sense is this noun class morphology as opposed to just a diminutive suffix? Agreement, presumably. There is, as I understand it, a system vaaaaguely like this in some native American languages' classificatory verb stems - I think you can, for example, use the classificatory stem for... I don't know... flimsy things? To imply that a man is drunk. There's also an African (?) language, I seem to remember, where you can refer to something that would normally be masculine with feminine gender and imply that it is fatter and rounder than a typical specimen of that class. It seems to me though that at least for the duration of that conversation, or whilst that constituent remains salient, the same agreement morphology would probably be used, rather than it being used once (essentially accompanying an adjectival modifier) and then dropped later on in the conversation. Also, these things obviously lend themselves very easily to semantic drift, and I would expect that even in a language that allowed this as a general rule, there would be plenty of pairs where there would be a lexical distinction.

[Mednij]

In what sense is this noun class morphology as opposed to just a diminutive suffix? Agreement, presumably. There is, as I understand it, a system vaaaaguely like this in some native American languages' classificatory verb stems - I think you can, for example, use the classificatory stem for... I don't know... flimsy things? To imply that a man is drunk. There's also an African (?) language, I seem to remember, where you can refer to something that would normally be masculine with feminine gender and imply that it is fatter and rounder than a typical specimen of that class. It seems to me though that at least for the duration of that conversation, or whilst that constituent remains salient, the same agreement morphology would probably be used, rather than it being used once (essentially accompanying an adjectival modifier) and then dropped later on in the conversation. Also, these things obviously lend themselves very easily to semantic drift, and I would expect that even in a language that allowed this as a general rule, there would be plenty of pairs where there would be a lexical distinction.

OK, in that case I was mistaken about a single noun belonging to multiple noun classes in a single conversation. So in your opinion such a system would be unstable?

[Ser]

Many nouns in Chinese can take more than one classifier, stressing different qualities about the nouns with each classifier.

Some examples given in Matthews and Yip's Cantonese: A Comprehensive Grammar are 佢把口 kéuih bá háu 'her mouth (as an instrument for speaking)' (把 bá is a classifier used with tools like forks and locks) vs. 佢個口 kéuih go háu 'her mouth (as a part of the body)' (個 is the generic classifier); 㗎船 ga syùhn 'ship (as a large vehicle)' (㗎 ga is a classifier for vehicles), 隻船 jek syùhn 'ship (as a small object)' (隻 jek is a classifier generally used for small animals like sparrows). However, such distinctions are not used to contrast two similar objects the way you described them in your second-to-last post. Classifiers are not that open either, as far as I know.