Talk:Relative clause

"old" and "new" analyses

The standard analysis of relative clauses as I learned it separates the strategies into two classes, one describing the relative pronoun or other means of joining the two clauses, and the other describing how the role of the shared noun is indicated in the embedded clause. This is what is found (AFAIK) in the standard textbooks on linguistic typology (e.g. Bernard Comrie, Jae Jung Song, etc.). I don't know where the existing ("old") analysis in this page comes from, which describes four main types with no separation of the two strategies. I think it's clearer in any case to separate them, and so I added this analysis (the "new" one) as the first one described. I left the "old" one in place afterwards, esp. since it contains some interesting details. But eventually this should be fixed up. Benwing (talk) 07:41, 30 August 2011 (UTC)

BTW, I think some fact checking may be needed of the "old" analsis. For example, it says (or rather, said) this:

Relative pronouns in the strict sense are almost entirely confined to European languages, where they are widespread except among the most conservative Celtic family.

First, I have a hard time believing that relative pronouns are really almost missing in non-European languages. Certainly, Classical Arabic has a relative pronoun (although it may not be "strict" in the sense that there is no case marking of the embedded role on the relative pronoun — there is only case marking on the dual, and it agrees with the role in the matrix clause; but, spoken English also has no case marking of the relative pronoun, since few people say "whom" and hardly any of those use it in a prescriptively "correct" way, Latin-style). Secondly, Celtic is hardly "most conservative" among IE languages (hence, I removed this peacock term entirely). Third, other branches lack Latin-style relative pronouns (e.g. the Indo-Aryan languages, with correlative pronouns). Benwing (talk) 07:51, 30 August 2011 (UTC)

Bc bc of my health insurance insurance 2001:4C4C:19DF:4600:5C15:81AE:8B3C:9852 (talk) 10:14, 1 January 2023 (UTC)

Errors in gloss of Chinese example

In one of the examples for the section on Chinese, there seems to be a few words missing in interlinear glosses (i.e. in the example with "（用）今天赢来的钱来付房租"). Here, "来" is matched with fu4 and "pay," when those should actually be under "付." The "来" could actually just be removed entirely; it seems to me that it may have been added by an editor, who forgot to doublecheck it with the other rows of the grid, and thus creating this error. Another mistake is that in the Chinese version, "房租" is together, while in the English annotations "房" and "租" are separate ("house" and "rent").

Also, I believe that the translation is a bit off; the translation given in the article is "'the won-today money pays the rent' or 'the money that was won today pays the rent,'" though this is nonsensical; a more accurate and logically correct translation (in the case where 用 is present) would be "use won-today money pay rent" or "to use the money that was won today to pay the rent." This brings me to my next point: here, "用" is necessary. Without it, the phrase is just an object with a verb. If one were to translate it, they would probably translate it as "the money that was won today pays the rent," which, well, happens to be what is written in the article, and like how I said above, it's a nonsensical phrase. — Preceding unsigned comment added by Your + 're (talk • contribs) 02:16, 1 June 2022 (UTC)

Lehmann citation is wrong/misleading

The citation for Lehmann's Der Relativsatz is:

Lehmann, Christian (1984). Der Relativsatz [Relative Clauses]. Language universals series; vol. 3 (in German). Tübingen: G. Narr. p. 438.

In fact, p. 438 is the very last page of Lehmann's book. So when the article says:

In other languages, relative clauses may be marked in different ways: they may be introduced by a special class of conjunctions called relativizers, the main verb of the relative clause may appear in a special morphological variant, or a relative clause may be indicated by word order alone.

and cites Lehmann for this information, there is no way of finding where in the 438 pages of this book the information is given. Unless Darigon Jr., who added the source, can provide pages on which this information is to be found, I will delete the source from the article. Bathrobe (talk) 19:27, 20 February 2024 (UTC)

Would you consider replacing it with a different citation? WALS agrees with the information, citing Lehmann 1984: https://wals.info/chapter/s8 and https://wals.info/chapter/122. Another candidate: https://wals.info/chapter/90 Femke 01 (talk) 15:24, 21 February 2024 (UTC)

The paragraph gives three examples of how relative clauses are "marked":

1. Introduced by a special class of conjunctions called relativizers.

2. The main verb of the relative clause may appear in a special morphological variant.

3. A relative clause may be indicated by word order alone.

These three are not properly followed up or illustrated in the rest of the article, nor are they properly sourced (the reference to Lehmann notwithstanding).

Your first WALS link refers to Lehmann's "prenominal", "postnominal", "circumnominal" classification. It also refers to "relative pronoun" and "gapping". This is insufficient to justify the content of the paragraph.

Your second WALS link refers to how the relativised position is indicated inside the relative clause ("relative pronoun", "nonreduction", "pronoun retention"). This does not provide a basis for the content of the paragraph.

The third WALS link is a classification of types, again focussing on "prenominal", "postnominal", "circumnominal" ("internally-headed"), as well as non-embedded types. It does add an (in the context) rather tangential note "that languages vary as to whether relative clauses involve forms of verbs that also occur in main clauses", which covers the main verb of the relative clause may appear in a special morphological variant.

None of the three links explicitly justify the statement that a relative clause may be indicated by word order alone.

The paragraph in question, which was present before the vague Lehmann reference was inserted by user Darigon Jr, gives no examples to illustrate what it is talking about and does not appear to represent any particular classification. If it were properly sourced and illustrated with examples it would be a useful and interesting part of the article. As it stands, it is simply a random, slightly baffling, essentially unsourced collection of suggestive statements that doesn't lead anywhere. Bathrobe (talk) 19:22, 21 February 2024 (UTC)

The paragraph in question was inserted by user Ruakh on 22 April 2005 with the explanation: "Added detail to first paragraph (should reduce non-NPOV)". It went:

"In other languages, relative clauses may be introduced by a special class of conjunctions called relativizers, or may be indicated by word order alone. In some languages, more than one of these may be possibilities."

On the same day, user MarkSweep elaborated on this as follows:

"In other languages, relative clauses may be marked in different ways: they may be introduced by a special class of conjunctions called relativizers; the main verb of the relative clause may appear in a special morphological variant; or a relative clause may be indicated by word order alone. In some languages, more than one of these mechanisms may be possible; and others yet do not employ relative pronouns or relativization marks at all."

Again on the same day, user Pablo-flores asked this:

"Is this so really? There should be examples of that".

User Ruakh responded: - "Yes, and there are examples of that; see #Japanese. (At least, I assume that's what the person who wrote that sentence meant. I suppose it's possible that he meant that you can't even identify relative clauses in some languages, but that seems unlikely.)"

It was much later that user Darigon Jr. added a reference to Lehmann, without indicating where in Lehmann this information was given.

Since it is an unsourced fragment inserted at a very early stage in the history of the article, and appears to have been kept purely through inertia, despite a huge amount of material being added and English relative clauses being moved to an article of its own, I suggest that the paragraph should be entirely deleted. Bathrobe (talk) 21:01, 21 February 2024 (UTC)

Or expanded properly to take account of the rest of the article. Bathrobe (talk) 21:02, 21 February 2024 (UTC)

Exactly, since the information is correct and relevant. There are lots of sources for examples if you google relativization typology. Borrowing from WALS saves time, but it may be over-cited already. Femke 01 (talk) 08:20, 22 February 2024 (UTC)

I do not understand why you say the information is correct and relevant, after I specifically pointed out how poor it is. It is a skimpy leftover from a much earlier article. It is barely relevant to this article (it was included to counterbalance a heavy bias towards English -- no longer relevant since English now has its own article) and fails to summarise anything much at all. The article itself dutifully presents various aspects of relative clauses, but it is poorly structured and reads like a confusing patchwork. This little paragraph doesn't summarise anything, is confusing, and the so-called "source" is an ad hoc addition. From whatever angle you look at it, it is a poor excuse at classifying relative clauses and would be better left out. A decent summary (and if you actually read the literature, not just WALS, you might find how hard it would be to create one) would be far more useful. Bathrobe (talk) 10:33, 22 February 2024 (UTC)

But if you just delete half of the introduction, the article will appear even more like a patchwork. I hear what you say and I agree that the article could be improved (as it is always) but I am unbiased in terms of not having contributed to it. Nonetheless, relativization is in my field of expertise or what I have studied, and to me the information seems fine even if the same might be expressed differently. Lehmann 1984 is also a classic authoritative source. I will not be stopping you if you delete it, though, so it is up to you. Femke 01 (talk) 14:00, 22 February 2024 (UTC)

if you just delete half of the introduction, the article will appear even more like a patchwork.

Actually, the first problem is the opening definition, which is from a grammar of English. There are other definitions that try (not always successfully) to cover relative clauses from a wider linguistic viewpoint, including (for instance) Hindi and Walbiri (Warlpiri).

A quick and incomplete note on a few random ways of indicating relative clauses (namely the use of relativisers, special verb forms, and, er, "word order", which appears to mean that sentences are just thrown into prenominal position) is a poor addition to the lede. It is also clearly referring to the familiar type of externally-headed, embedded relative clause and lacks a wider perspective.

As to the source, of course Lehmann is a classic. But to be properly sourced you need page numbers. Lehmann is very long. Where exactly did he provide a basis for what is stated in that paragraph?

On both of these counts this paragraph is sorely lacking. I do not think it is salvageable in its current form. In fact, the entire lede really needs to be rewritten and properly sourced.

Bathrobe (talk) 18:29, 22 February 2024 (UTC)

True. ~~~~ Femke 01 (talk) 18:59, 22 February 2024 (UTC)

Section structure confuses me: role in higher-ranking and in lower-ranking clause

The current article contains:

”English can relativize all positions in the hierarchy. Here are some examples of the NP and relative clause usage from English: […] Languages that cannot relativize directly on noun phrases low in the accessibility hierarchy can sometimes use alternative voices to "raise" the relevant noun phrase so that it can be relativized.”

The section containing this text begins with focussing on the role in the higher-ranking clause and then shifts focus to that in the lower-ranking clause. Even the introduction to the examples in […] seems to refer to the role in the higher-ranking clause.

But … the first groep of examples does not seem to vary with respect to the role of the relativized entity in the higher-ranking clause. Rather, it seems to focus om the role in the lower-ranking clause. That confuses me.

Only the second group of examples in […] seems to focus on role differences in the higher-ranking clause.

I think the text could do with a more explicit distinction between the clause rank the roles in which are discussed.Redav (talk) 10:21, 6 April 2024 (UTC)

Correction: Also the second group focusses on therole in the lower-ranking clause.

Addition: I was expecting examples like:

1. The A, who is standing, gives him the B of the C for the D.
2. The A gives him, who came first, the B of the C for the D.
3. The A gives him the B, which is lying on the table, of the C for the D.
4. The A gives him the B of the C, who looks unhappy, for the D.
5. The A gives him the B of the C for the D, whom we are now looking for.

One might think of:

“The man gives him the book of the writer for (i.e. on behalf of; as a favour to) the mayor.”

I would say this covers the roles of subject, indirect, direct object, genitive and oblique object / benefactive in the higher-ranking clause.Redav (talk) 10:47, 6 April 2024 (UTC)

The article uses the wording “to relativize (on)”. I have understood the direct (or prepositional) object of this verb as the entity in the higher-ranking clause that is referred to in the lower-ranking clause. If, conversely, the intended object is the entity in the lower-ranking clauses that refers to an entity in the higher-ranking clause, the use of the wording “to relativize” confuses me.Redav (talk) 10:55, 6 April 2024 (UTC)

I think I can now see my mistaken understanding of “to relativize”.

Nevertheless, because of the structuring of the section, I still expected examples similar to the ones I have provided as well as and preceding the examples such as are in the article text.Redav (talk) 11:04, 6 April 2024 (UTC)

Gapping in languages with relative pronouns or adverbs, and pragmatic word order?

Can you really say that such languages (like German, or even stronger Classical Latin) do employ gapping? Since these languages do include the argument in question in the relative clause by employing the relative pronoun or relative adverb, and there is no fix position for that argument in the remaining parts of the clause, I do not see where should be a gap in such clauses. 2A0A:A541:F78F:0:D830:148B:C1C7:B983 (talk) 15:00, 22 May 2024 (UTC)

I think that is an excellent question but I don't know how to answer it. You are obviously familiar with the linguistic literature on this so I won't try to lecture you on basics. I will simply quote from a paper by Thilo Weber (The Proform-Conjunction Interface: A Study of the Syntax of Relative That), which presents a description of English relative clauses in terms of gapping. And since linguistic theory is (de facto) usually based on English, that is probably why this analysis is usually adopted.

Within the framework of GB, the formation of bound relative clauses is a process that involves deletion of the relativised constituent, insertion of a wh-constituent, and movement from its logical D-structure into clause-initial position, where it appears in S-structure. ..... The moved constituent is considered to leave a "gap" or "trace" in its original position.

If, as you point out, Classical Latin represents the argument in the relative clause through the use of the relative pronoun or relative adverb, and there is no fixed position for that argument in the clause, then the concept of gapping is possibly in question. But if there is any kind of movement of the relative pronoun within the relative clause -- which grammatical theory would have to account for -- then the gapping treatment would still be valid (within generative theory, at least).

I would like to see someone else comment on this. Empty elements are rife in generative linguistics (if in doubt, just move it and leave a gap behind), but that is a theoretical issue that I am not willing or qualified to tackle.

(I am only commenting here because I think your question raises a reasonable doubt, which no one else has responded to.)

Bathrobe (talk) 22:56, 26 August 2024 (UTC)