What are the categories of pronouns?

[HolyKnowing]

What are the categories of pronouns in languages, starting with English as a base?

Here are some that come to mind:

1. Stand-alone pronouns, which are pronouns that refer to people, or objects connected to people. {I, You, He, It, etc...}. It is a popular paradigm to construct personal pronouns according to person and number, usually resulting in six personal pronouns, generally speaking. Most European languages augment the 3rd person singular with gender divisions. Hebrew augments 4/6 of them with gender divisions. Tok Pisin augments 1st person plural with clusivity. Really traditional English speakers endow honorifics to first person plural ("royal we") and second person singular ("thou").

Japanese has a pronoun system not based on a paradigm such as person × number, but based on the sociology of groups in Japan.

Possessives are polymorphs of personal pronouns, therefore they don't receive special treatment.

2. Binding pronouns. These are dependent upon stand-alone personal pronouns. Reflexive forms, such as himself, count. Reciprocal forms, such as each other, also count. Some languages like Norwegian have more binding pronouns. Other languages such as Farsi, I believe, also have resumptive binding pronouns for relative clauses.

3. Relative pronouns. These deserve a category but they're so rich that treatment of them will be deferred to a future time. There's some intersection with binding pronouns (because of the linguistic possibility of resumptive binding pronouns), which means the category of pronouns cannot be represented by something as simple as a lattice. Although English's

4? A hypothetical category of pronouns are component pronouns, which refer to the parts within a sentence. An example of a hypothetical category of component pronouns would be "Jack London fantasizes about being an animal, just like in book Call of the Wild by the subject.)" However no language I know of has pronouns that refer to parts within the sentence, such as the subject of the sentence.

Any thoughts, reflections, and connections?

[Travis B.]

Really traditional English speakers endow honorifics to first person plural ("royal we") and second person singular ("thou").

Actually, you was honorific, and thou was familiar. It just turned out that thou became subject to pejoration except when used for prayer and other religious contexts, and hence got replaced with you in most language outside of religious contexts. Consequently, people started perceiving thou as honorific once it had largely fallen out of use, as it was perceived as religious in nature.

[Creyeditor]

What about indefinite pronouns? And wh-pronouns?

[bradrn]

Bhat’s Pronouns (2004) is the best reference I know on this subject. He doesn’t actually give a single list, but reading through the book, he covers the following types of pronouns:

• Personal pronouns:
  • Free pronouns (of the sort you call ‘stand-alone’):
    • first/second-person pronouns
    • third-person pronouns (which very often have different status)
  • Bound pronouns (variously called ‘agreement’, ‘cross-referencing’ etc.)
  • Pronouns which specify coreference: logophoric pronouns, anaphoric pronouns (including reflexives, as well as so-called ‘long-distance’ reflexives)
• Pro-forms (i.e. non-personal pronouns)
  • Demonstratives
  • Indefinites and interrogatives (often merged)
  • Relative and anaphoric forms (again, often merged)

Note that personal pronouns and pro-forms are very different in their formation and behaviour, and should not be treated as the same thing.

Japanese has a pronoun system not based on a paradigm such as person × number, but based on the sociology of groups in Japan.

Many linguists argue that Japanese doesn’t have personal pronouns at all — these are merely nouns. (This also applies to Burmese, Thai and probably other languages of the region too.)

[zompist]

Many linguists argue that Japanese doesn’t have personal pronouns at all — these are merely nouns. (This also applies to Burmese, Thai and probably other languages of the region too.)

This is a complicated issue, but I always feel this position rests on a misconception: that all pronouns should work like Latin, being unanalyzable roots without alternatives.

Japanese works almost entirely like Portuguese: there's a noun-to-pronoun pipeline.

Portuguese inherited tu and vós from Latin, but these are hardly used. In medieval times you said vossa mercê 'your mercy'. This became vosmecê and then você and then, in very colloquial speech, cê. Você is used exactly like (say) French vous; is it not a pronoun? It derives from an NP, obviously, but it no longer means what it did. If you acuallly want to say 'your mercy' ("Your mercy and kindness are admirable") you can't abbreviate it você.

Or, more pithily, in colloquial French nous is replaced with on. This derives from Latin homo 'person', but it's a pronoun now.

For formal 'you', Portuguese often uses o senhor or a senhora. These are somewhere in the middle of the pipeline.

You can also, as in English, use expressions like "your excellency", "your majesty". Nobody really calls these pronouns, but they're also not exactly regular NPs. They are highly conventionalized and often stretch the morphology. (E.g. Spanish su majestad is feminine gender, but (to my understanding) you'd normally say Su majestad es muy sabio, "Your majesty(f) is very wise(m)."

Now let's look at Japanese. It's complicated by being largely pro-drop, so often you don't use pronouns at all; but that's true of Spanish too.

It's also complicated because you use titles for superiors, including within the family. You call your older sister onēsan, your teacher sensei, etc. These are simple nouns, but they're used where English speakers would usually use pronouns. (But not always: nothing weird about calling your father Dad.)

Then we come to what a Japanese textbook will call pronouns— watashi, boku, anata, atashi, etc. These almost always have etymologies as nouns, but etymology is not current grammar! E.g. boku derive from 'servant', but it does not mean that any more; in fact it's considered a masculine and even bold way of referring to oneself. If you actually have a servant they are a meshitsukai (or other terms), not a boku. Over the centuries, Japanese pronouns start out humble, get reinterpreted as too familiar, and are replaced. This isn't much like English, but it's not really different from Portuguese (where e.g. você is now too familiar for a stranger and is replaced by o senhor.

[bradrn]

These almost always have etymologies as nouns, but etymology is not current grammar!

This is the really key question, I think — in current grammar, how do Japanese ‘pronouns’ behave? I think I’ll just quote Bhat in full here, since he lists many good syntactic tests along these lines:

It appears to me that a definite answer to this question can be provided only through an examination of the use of those nouns […] That is, we need to find out answers to several questions of the following type. Do these languages require noun phrases to be used in matrix sentences, in addition to the [ostensible] personal pronouns, for identifying the speaker and the addressee […]? Do these languages allow the nominal expressions that occur as personal pronouns to take modifiers and complements such that they can establish the identity of the participants by themselves? Do they require noun phrases to be placed in apposition in an oath-taking context […]? Depending upon the answers that we obtain to these questions, we would be able to say whether the nominal expressions that occur in the place of personal pronouns in these languages are to be regarded as personal pronouns or merely as nominal expressions.

It is possible, however, that these nominal expressions show a gradation in manifesting the differentiating characteristics of personal pronouns in these languages. Sugamoto (1989) makes such a claim regarding the personal pronouns of Japanese, based upon an examination of some of their characteristics. According to Hinds (1986: 241) also, personal pronouns of this language differ from the pronouns of other languages like English (i) in having nominal origins, (ii) in being terms of occupation or status titles, (iii) in being very large in number, with different forms being selected depending upon sex, age, perceived social status, and emotional correlation, and (iv) in showing most of the nominal characteristics like occurring after demonstratives and being modified by adjectives or relative clauses. Hinds concedes, however, that the nominal expressions show some characteristics of their own, which makes it necessary to regard them as forming a distinct 'pronominal' category.

(Yes, he has a somewhat ponderous style… I’ve done my best to condense it. It’s easier to read in the typeset book, for some reason.)

[Travis B.]

You can also, as in English, use expressions like "your excellency", "your majesty". Nobody really calls these pronouns, but they're also not exactly regular NPs. They are highly conventionalized and often stretch the morphology. (E.g. Spanish su majestad is feminine gender, but (to my understanding) you'd normally say Su majestad es muy sabio, "Your majesty(f) is very wise(m)."

Of course, consider NAE you guys, where guy originates in the noun guy... which originates in the first name of Guy Fawkes of all people...

[HolyKnowing]

What about indefinite pronouns? And wh-pronouns?

Really good question. The answer is that many pronouns aren't actually pro-forms but rather means of encoding grammar (wh-pronouns have dual functions of constructing relative clauses and questions) or logic (some, all, much, many, and the "table of correlatives"-type of pronouns serve as a way of representing logic).

[bradrn]

What about indefinite pronouns? And wh-pronouns?

Really good question. The answer is that many pronouns aren't actually pro-forms but rather means of encoding grammar (wh-pronouns have dual functions of constructing relative clauses and questions) or logic (some, all, much, many, and the "table of correlatives"-type of pronouns serve as a way of representing logic).

I’m not quite sure what you’re trying to say here, but I don’t understand it. Firstly, ‘encoding grammar vs encoding logic’ doesn’t strike me as a meaningful distinction. Secondly, wh-pronouns (i.e. conflated relative/interrogative pronouns) and and indefinite pronouns fit quite comfortably within the ‘table of correlatives’, in languages which have such a system.

[HolyKnowing]

Firstly, ‘encoding grammar vs encoding logic’ doesn’t strike me as a meaningful distinction.

You understood correctly. I was separating (causative of segregate, or is segregating the middle voice of separating?) pro-forms from the remaining pronouns, the remainder being halved into "pronouns encoding grammar" and "pronouns encoding logic".

Secondly, wh-pronouns (i.e. conflated relative/interrogative pronouns) and and indefinite pronouns fit quite comfortably within the ‘table of correlatives’, in languages which have such a system.

You understood correctly. Wh-pronouns and indefinite pronouns form a nice pile with the remaining correlatives.

[bradrn]

Firstly, ‘encoding grammar vs encoding logic’ doesn’t strike me as a meaningful distinction.

You understood correctly. I was separating […] pro-forms from the remaining pronouns, the remainder being halved into "pronouns encoding grammar" and "pronouns encoding logic".

I’ll repeat my question: if you’re halving personal pronouns into those two categories, then what do you consider to be the difference between those two categories?

(causative of segregate, or is segregating the middle voice of separating?)

I don’t know what your native language is, but English has no middle voice and no productive morphological causative. ‘Segregate’ and ‘separate’ are simply near-synonyms.

[xxx]

in 3SDL, each semantic prime constituting a concept
can play the role of pronoun by anaphora outside this concept...

[HolyKnowing]

I’ll repeat my question: if you’re halving personal pronouns into those two categories, then what do you consider to be the difference between those two categories?

Grammar is something internal to the language, so pronouns that create new grammar like wh-pronouns are internally-oriented. Logic is something external to the language, so pronouns that describe logic are externally-oriented.

I don’t know what your native language is, but English has no middle voice and no productive morphological causative. ‘Segregate’ and ‘separate’ are simply near-synonyms.

Causative and middle voices are concepts that exist independently of any particular language, which is what makes it possible for linguistics to exist as an intellectual field of study. This is why causative and middle voice are applicable to many verbs and verb phrases in the English language. For instance, you can interpret "to lose [the subject's] temper" as the middle voice of "to anger". Or you can interpret "to nurture" as the causative of "to grow".

[bradrn]

I’ll repeat my question: if you’re halving personal pronouns into those two categories, then what do you consider to be the difference between those two categories?

Grammar is something internal to the language, so pronouns that create new grammar like wh-pronouns are internally-oriented. Logic is something external to the language, so pronouns that describe logic are externally-oriented.

What does it even mean, in your opinion, for something to be ‘external to language’? The distinction you’re trying to draw here makes absolutely no sense to me.

I don’t know what your native language is, but English has no middle voice and no productive morphological causative. ‘Segregate’ and ‘separate’ are simply near-synonyms.

Causative and middle voices are concepts that exist independently of any particular language, which is what makes it possible for linguistics to exist as an intellectual field of study. This is why causative and middle voice are applicable to many verbs and verb phrases in the English language. For instance, you can interpret "to lose [the subject's] temper" as the middle voice of "to anger". Or you can interpret "to nurture" as the causative of "to grow".

This is basically a quibble about what we mean when we say words like ‘causative’. In one sense, yes, there is a general notion of ‘causing’, which is essentially a language-independent category. But languages grammaticalise this notion to different extents. In English, we have plenty of causative verb pairs, such as ‘rise’ vs ‘raise’. However, English doesn’t have many devices to systematically produce causative verb forms: this being why I said, specifically, that English has ‘no productive morphological causative’.

(‘Nurture’ vs ‘grow’ isn’t actually a great example here: if I ‘nurture’ something, that has additional connotations over simply ‘making something grow’.)

‘Middle voice’ is a more difficult one. Certainly, many languages are said to have some morpheme called a ‘middle’, and those morphemes display some similarities. But are these varied ‘middles’ really all instance of one single semantic relationship? Some people have argued that they are, but others have argued that they aren’t. Personally, I suspect that they aren’t, and the so-called ‘middle voices’ are really language-specific conflations of more universal detransitivising categories (anticausatives, reflexives, reciprocals, etc.).

A further issue here is that I’ve actually never heard people use ‘middle’ as a semantic relationship between lexical verbs (as ‘causative’ can be used). Instead, it invariably refers to a construction which systematically derives intransitive verbs from basic transitive verbs. This is probably another sign that ‘middle’ doesn’t refer to a single, cross-linguistically valid idea.

(In fact, the construction usually called the ‘English middle’ is something quite different to your examples: people use the term of constructions like ‘the book sells well’, where an active verbs takes what appears to be an object. For your example, it’s probably more justifiable to call the relationship a causative one between ‘anger’ and ‘get angry / lose one’s temper’.)

EDIT: just managed to find a good paper on the subject of the middle voice and its semantics: https://doi.org/10.1515/lingty-2020-0131

[zompist]

[This is the really key question, I think — in current grammar, how do Japanese ‘pronouns’ behave? I think I’ll just quote Bhat in full here, since he lists many good syntactic tests along these lines:

It is possible, however, that these nominal expressions show a gradation in manifesting the differentiating characteristics of personal pronouns in these languages. Sugamoto (1989) makes such a claim regarding the personal pronouns of Japanese, based upon an examination of some of their characteristics. According to Hinds (1986: 241) also, personal pronouns of this language differ from the pronouns of other languages like English (i) in having nominal origins, (ii) in being terms of occupation or status titles, (iii) in being very large in number, with different forms being selected depending upon sex, age, perceived social status, and emotional correlation, and (iv) in showing most of the nominal characteristics like occurring after demonstratives and being modified by adjectives or relative clauses. Hinds concedes, however, that the nominal expressions show some characteristics of their own, which makes it necessary to regard them as forming a distinct 'pronominal' category.

I'd note that he's saying a lot of the same things I said, down to the idea of a continuum (his "gradation").

Does he give examples of demonstratives used with personal pronouns like watashi? That sounds interesting.

Pronouns can take adjectives in English: foolish me! I gave a Latin example in the SCK. This may be rare, but it's not a bright-line distinction from Japanese. Again I'd like to see Japanese examples.

What does he mean about relative clauses? "You who have served the country deserve our gratitude. He who laughs last laughs most. They who must not be named caught me."

I'd also love to know what Hinds is talking about ("some characteristics of their own"), for the next time this comes up!

[Linguoboy]

Now let's look at Japanese. It's complicated by being largely pro-drop, so often you don't use pronouns at all; but that's true of Spanish too.

It's also complicated because you use titles for superiors, including within the family. You call your older sister onēsan, your teacher sensei, etc. These are simple nouns, but they're used where English speakers would usually use pronouns. (But not always: nothing weird about calling your father Dad.)

Korean differs from Japanese in that it possesses "true" first- and second-person pronouns without obvious nominal etymologies[*], but I think the usage is similar insofar as the use of these pronouns is much more restricted than in SAE in general and English in particular. It's not just that it's pro-drop but that in many case where SAE would use an anaphoric pronoun to clarify reference, Korean would use instead a title, kinship term, or given name. I mainly associate liberal pronoun usage with formal and/or English-influenced registers (which arguably includes contemporary pop music).

[*] Third-person pronouns are an entirely different matter, being a combination of demonstratives with bound nominal morphemes or sometimes just bare demonstratives.

[bradrn]

I'd note that he's saying a lot of the same things I said

Perhaps I misread your post, then; he struck me as being a little more specific.

Does he give examples of demonstratives used with personal pronouns like watashi? That sounds interesting.

Pronouns can take adjectives in English: foolish me! I gave a Latin example in the SCK. This may be rare, but it's not a bright-line distinction from Japanese. Again I'd like to see Japanese examples.

What does he mean about relative clauses? "You who have served the country deserve our gratitude. He who laughs last laughs most. They who must not be named caught me."

I'd also love to know what Hinds is talking about ("some characteristics of their own"), for the next time this comes up!

The reference is simply entitled Japanese, from the Croom Helm series. And, amazingly, I actually have a copy of the book! So let’s see…

So, starting with demonstratives (as he does), here’s his example:

dakedo,

but

ima,

now

[sono

that

kanojo]

she

mo

too

ne,

EM

nanka,

like

kookai,

regret

kono

this

mae

before

tegami

letter

ni

in

kaite

writing

kite

coming

ne,

EM

yappari

as-expected

kookai

regret

shite

doing

ne

EM

However, that she has written [to me] telling [me] that [she] regrets, as expected, that [she] regrets [what she did].

Sadly, he doesn’t discuss what this means or when this is used.

With relative clauses, his next example is:

[kekkon

marry

shite

do-and

juunanasai

17-years

de

by

kodomo

child

o

AC

unda

bore

mono

even

no

LK

sodatete-iku

raise

nooryoku

ability

ga

NM

nakatta]

NEG-past

kanojo

she

wa

TP

sansai

3-years

no

LK

toki

time

danii

Danny

o

AC

yooshi

adopted-child

ni

to

dashita

sent-out

She who did not have the ability to raise a child even though she bore one at the age of 17 after getting married put Danny out for adoption when [he] was three years old.

Again, he doesn’t discuss this properly, but it strikes me as being less restricted than the English equivalent.

However, the really interesting part comes with the next examples. In his words, these show how pronouns can ‘serve also as nicknames or terms of address, a feature which is very nominal in nature’:

boku,

I

doo

how

shita

did

no?

QU

Boku [ie, little boy], what happened?

oi

hey

kanojo!

she

sake

sake

motte

bring

koi!

come-imperative

Hey kanojo [ie, you]! Bring some sake! (rough and impolite)

To me, this seems like quite strong evidence that these are distinctly closer to nominal status than is the average ‘pronoun’ category.

[Richard W]

4? A hypothetical category of pronouns are component pronouns, which refer to the parts within a sentence. An example of a hypothetical category of component pronouns would be "Jack London fantasizes about being an animal, just like in book Call of the Wild by the subject.)" However no language I know of has pronouns that refer to parts within the sentence, such as the subject of the sentence.

There are reflexive pronouns that refer to the subject of the sentence, best known to me from the older Indo-European languages, like Latin se, but still surviving in several modern languages. Now, the reflexive pronouns in s- are mostly 3rd person, but I believe there are some languages where they can be used for any person, as seen in the Russian and Swedish passives.

[Richard W]

English seems to have a rather peculiar pronoun, 'one', which takes the place not of a noun-phrase or equivalent, but simply a noun, as in, "I drive a blue car, but I'd rather have a red one." I'm not aware of an equivalent in any other language. Strikingly, this pronoun seems to be becoming a pronominal base for possessive pronouns - some people seem to regularly say "my one" for "mine". The nearest parallel I can think of is the use of classifiers in Thai, where it seems to commoner to form the equivalent of demonstrative pronouns from classifier + deitic adjective rather than to switch the tone on the adjective*. (Definite noun phrases with deitic adjectives also normally use the classifier.)

*One's not allowed to classify this as an inflection.

[Travis B.]

English seems to have a rather peculiar pronoun, 'one', which takes the place not of a noun-phrase or equivalent, but simply a noun, as in, "I drive a blue car, but I'd rather have a red one." I'm not aware of an equivalent in any other language. Strikingly, this pronoun seems to be becoming a pronominal base for possessive pronouns - some people seem to regularly say "my one" for "mine". The nearest parallel I can think of is the use of classifiers in Thai, where it seems to commoner to form the equivalent of demonstrative pronouns from classifier + deitic adjective rather than to switch the tone on the adjective*. (Definite noun phrases with deitic adjectives also normally use the classifier.)

*One's not allowed to classify this as an inflection.

The thing I notice is that when "one" is to used to refer to nouns other than people, it often implies some kind of contrast between the noun it refers to and some other noun of the same kind. For instance, "my one" implies that there is "one" which is not "mine", unlike "mine" itself.