Why is "that" in the relative clause not a complementizer?

Question

English is a very beautiful and systematic language. English is not my mother tongue. I have been interested in deep grammar for some time and I am confused.

Example: The girl that I love.

They say that here (in all of them) that is a relative pronoun, but my logic doesn't get it.

Here's what I think:

The girl that I love whom. (you will say that there is no such usage but my logic says that it has to be)

Isn't that above a complementizer?

Then the sentence goes like this:

The girl who that I love.

This time, either who or that is removed from the sentence.

The sentence becomes

The girl who (that hidden) I love or The girl (rp hidden) that I love.

Here that was always in the same place, even when there was no relative pronoun.

Now I wonder, why is that a relative pronoun? Isn't that a complementizer?

Forgive my ignorance, could it be that that, which is a relative pronoun, has been perceived as a relative pronoun over time?

Answer 1

In English, the word that is never a relative pronoun. This is the view taken by Huddleston & Pullum (2002); although I sometimes find their takes too galaxy-brained, I think that they're just correct on this point and that the traditional analysis is wrong. I'm going to start by explaining their view and then giving the arguments that support it (see pp. 1056-1057, 951-956).

In English, the word that is sometimes used to mark a finite declarative clause as subordinate. Consider these three sentences:

1. I know (that) she is smart.
2. That she is smart is very well-known.
3. I trust her because she is smart.

In (1), the marker is optional: you can mark "she is smart" as a subordinate clause explicitly by adding the "that," but doing so isn't required. In (2), the marker is obligatory: you need to mark "she is smart" as subordinate in sentences like that, and *"She is smart is very well-known" would be incorrect. In (3), the marker is forbidden; even though "she is smart" is a subordinate declarative clause, *"I trust her because that she is smart" is ungrammatical, since (essentially) you can't have both a "because" and a "that" together.

The same applies to relative clauses, since they are also subordinate declarative clauses. Let's consider these cases:

4. She is the person (that) I trust.
5. She is the person that trusts me.
6. She is the person who I trust.

In (4), as in (1), the that is optional; you can explicitly mark "I trust" as subordinate by adding a that, but doing so isn't required. In (5), as in (2), the that is mandatory; there are certain structural conditions in which subordinate clauses need a that, and (5) and (2) are both examples. In (6), as in (3), the marker is forbidden: even though "who I trust" is a subordinate declarative clause, you can't mark it with a "that." The sentence *"She is the person that who I trust" is ungrammatical, since (essentially) you can't have both a "who" and a "that" together.

It is true that, in (4) and (5), "I trust" and "trusts me" are not complete clauses on their own. As is typically the case, these relative clauses contain a gap, or missing element, that we generally understand as referring (or rather anaphoric) to the noun that the clause is modifying: "I trust ____" or "____ trusts me," where the "____" refers to the person in question. (There's also a gap in (6), in "[who [I trust ____]]," but there the gap is anaphoric, not to person, but to the word who, and it's the word who that is anaphoric to person.)

(I should note that Huddleston & Pullum don't themselves directly compare examples like (1)-(3) with (4)-(6), as I've done here, but I think this approach more clearly explains the view they endorse.)

So, how do we know that that isn't a relative pronoun? There are four reasons listed by Huddleston & Pullum (again, pp. 1056-1057). These are:

First: if that were a pronoun, it would have to be one that can be used much more widely than the uncontroversial relative pronouns who or which, or indeed than any other pro-form in the language. Compare sentences like: "He was due to leave the day that she arrived" and "He followed her to every town that she went." If we take that as a pronoun, we would need to maintain that it can mean not just who and which, but also where and when. Indeed, there are some cases where it isn't equivalent to any other relative pronoun, as in "She's the happiest that she has ever been."

Second: unlike the relative pronouns who and which, the word that cannot be part of a larger relative phrase, e.g. as the object of a preposition. "He's the person who I was referring to" can be turned into "He's the person to whom I was referring," but "He's the person that I was referring to" cannot turn into *"He's the person to that I was referring."

Third: unlike who/which, the pronoun "that" can only be used with finite declarative clauses, not with clauses of other kinds. "That's the knife with which to cut the vegetables" is correct, but neither "That's the knife with that to cut the vegetables" nor "That's the knife that to cut the vegetables with" is acceptable.

Fourth, and most importantly: just like the that that introduces subordinate clauses in (1)-(3), the word that is omissible under certain circumstances, but mandatory in cases where they are needed to mark the start of a certain clause due to certain structural conditions. No other pronoun in English behaves like this.

I would add one further point. The demonstrative that found in sentences like "I know that," which traditionally we'd consider a pronoun, is typically pronounced /ðæt/. But the word that found in sentences like (1) and (4) can be reduced to /ðət/, suggesting a further similarity between the two.

So: is that a complementizer? Probably not, at least if you assume that a complementizer has to be the head of the phrase containing it (see ibid. pp. 955-956). First, the fact that the word that is omissible in sentences like (1) and (4) suggests that "that" is not the head in phrases like "that she is smart" or "that I trust"; instead, that is best treated as a dependent of the clause following it. Secondly, some verbs require or allow a subjunctive clause instead of a declarative one; if that is the head, then we'd struggle to explain why a that clause containing a subjunctive is allowed in some places but not others. Finally, the word that can sometimes occur within the subordinate clause, as in "The boat was such an attraction that I was afraid[, if he came near it again, that I should never see the last of him]" (where the brackets indicate the subordinate clause).

Answer 2

The equivalent of a complementiser in grammars such as CGEL is a subordinator, and that's exactly the category that CGEL assign to relative that.

Relative that is also a complementiser in many other grammars, too.

However, there are a wide range of different analyses out there. In the other answers here, you'll find lots of arguments about why relative that should be regarded as a pronoun or something else. That doesn't alter the fact that many people share the Original Poster's educated opinion that it's a complementiser. Of course, you have to believe in complementisers to hold that view!

Answer 3

[1] The girl [that I love ___] lives in London.

[2] The girl [whom I love] lives in London.

[3] The girl [I love ___] lives in London.

"That" is not a relative pronoun, but a subordinator. It is the same "that" which introduces declarative content clauses, compare "I know [that you love me]".

In [1] there is no relative pronoun. "That" is a noun phrase functioning as prenucleus in the relative clause. A gap '___' is inserted to mark the position of the object of love", and this gap is anaphorically related to the antecedent "girl". The relative clause is marked as subordinate by the presence of the subordinator "that".

In [2] the relative clause is marked as subordinate by the presence of the relative pronoun "whom", which derives its interpretation from "girl", its antecedent.

[3] is a 'bare relative', so called because it lacks the subordinator "that" or a relative pronoun. The relative clause is marked as subordinate by the absence of the object of "love".

Answer 4

Alphabet’s answer has a compelling argument as to why that is not a relative pronoun. My answer still argues that it is one, but it may very well not be—certainly it is debatable.

TL;DR: That can function as both a relative pronoun and a complementizer, depending on the situation. In your case it is indeed a relative pronoun.

Both who and that can function as relative pronouns. Who (with objective form whom and possessive whose) is used for a human/personal antecedent, whereas that is usually used to head a restrictive clause with a non-human/non-living antecedent; for nonrestrictive clauses, which is used. (Note that in prepositional phrases, which is always used, and for both, the possessive is still whose.) You would never use who and that directly next to each other; pick one depending on what the antecedent is. For your example it would be

The girl whom I love. (Whom because it is the object of “I love”)

(While The girl that I love may be acceptable, who is universally so.)

If your antecedent were not a person, you would use that, as in

The chair that I love.

Or often the relative pronoun is omitted and simply implied if unambiguous:

The girl I love.

The chair I love.

There is another distinct use of that as a subordinating conjunction (i.e., complementizer) to introduce noun clauses, typically those following actions of speaking, thinking, observing, showing, etc. Here are some examples:

I saw that the sky was getting dark.

He knew that it was over.

She showed him that she could be trusted.

Again, that is often dropped from such sentences if doing so doesn’t hurt comprehension.

Answer 5

buIn this sentence you have two verbs and so you must have two clauses. It is fairly evident that "The girl that I love lives in London" is the main clause.

In "That I love", "that" has to be a relative pronoun because "to love" is a transitive verb, and so it must have an object in the clause; there is no other possibility than "that" for an object. Moreover this clause is a restrictive clause ("not any girl but the girl I love"); it is well known that only "whom", "that" and "zero pronoun" have this property as objective case pronouns. It would be complete nonsense to make of subordinators the object of the verb.

=================================================================== ADDITION

The text that is contested in CGEL has been copied in full below for those who would like to know exactly what is being criticized (CGEL's demoting "that" from the rank of relative pronouns).

3.5.6 That as a subordinator (not a relative pronoun)
Traditional grammar analyses the that which introduces relative clauses as a relative pronoun, comparable to which and who, but we believe that there is a good case for identifying it with the subordinator that which introduces declarative content clauses.
(a) Wide range of antecedent types and relativised elements
If that were a pronoun, or pro-form, its use would be much wider than that of the uncontroversial relative pronouns, or indeed of any pro-form at all in the language. Compare:
[71]
i They gave the prize to the girl [that spoke first]. [who]
ii Have you seen the book [that she was reading]? [which]
iii He was due to leave the day [that she arrived]. [when]
iv He followed her to every town [that she went]. [where]
v That’s not the reason [that she resigned]. [why]
vi I was impressed by the way [that she controlled the crowd]. [ ^** how]
vii It wasn’t to you [that I was referring]. [no wh form]
viii She seems to be the happiest [that she has ever been]. [no wh form]
It would not only cover the ground of all the simple ‘wh’ words put together, as shown in [i–v]: it would also appear in a variety of constructions where no ‘wh’ word could replace it, as in [vi–viii]. Particularly important here is the cleft construction shown in [vii], and in [70iii–iv] above. Note that, leaving aside the disputed case of the relative construction, there is no pro-form in English that takes as antecedent such complements and adjuncts as to you (in the sense it has in [71vii]), with considerable misgivings, in order to avoid this kind of misunderstanding, and the like. Instead of postulating a pro-form with such an exceptional range of use, we are saying that that relatives do not contain any overt pro-form linked to the antecedent: they simply have an anaphoric gap, like bare relatives.
(b) Lack of upward percolation
There are no that relatives matching wh relatives with a complex relative phrase:
[72]
i a. the woman [whose turn it was] b. the woman[that’s turn it was]
ii a. the knife [with which he cut it] b. the knife [with that he cut it]
If that were a pronoun we would have to stipulate that it has no genitive form, and that it never occurs as complement of a preposition – or rather that when it is complement of a preposition the latter must be stranded, for the knife that he cut it with is quite grammati- cal. The severe restrictions here stand in sharp contrast to the remarkable versatility of the putative pronoun that illustrated in (a). In the analysis where that is a subordinator the un- grammaticality of [72ib/iib] is predictable. Subordinators do not inflect and must occupy initial position; there is no relative word and hence no possibility of the relative feature percolating upwards into a larger constituent.⁹
(c) Finiteness
That relatives are always finite, as are the declarative content clauses introduced by that. Note, then, that we cannot insert that into non-wh relative infinitivals like a knife to cut it with – cf. ^*a knife that to cut it with. If that were a pronoun this would be a special fact needing explanation, but under the subordinator analysis it is exactly what we would expect, given that that is a finite clause subordinator.¹⁰
(d) Omissibility As we have noted, that can be regarded as very largely omissible in relative clauses in the same way as in declarative content clauses. The conditions under which omission is prohibited are not the same in the two cases (those for content clauses are given in Ch. 11, §3.1), but in both they have it in common that they are related to the need to mark explicitly the beginning of a subordinate clause under certain structural conditions. And in both cases, moreover, that is more readily omitted in simple structures than in complex ones. There is no pro-form in English that is systematically omissible under remotely similar conditions.

⁹ _{There are non-standard regional dialects of English in which that’s does occur, as in the man that’s leg was broken. We do not believe that such examples necessitate a pronoun analysis for the dialects concerned, and certainly they do not establish this analysis as valid for all dialects.}
¹⁰ _{The force of this argument is diminished by the fact that which can’t occur here either: we have a knife with which to cut it, not a knife which to cut it with. The absence of ^*a knife with that to cut it is then already covered under point (b). Nevertheless, the analysis of that as a finite clause subordinator does provide a very general account of why the only type of bare relative that can’t be expanded by means of that should be the infinitival one.}

---

(a), (b), (c), and (d) are considered below in some detail, but very probably, more remains to be said.

I (Wide range of antecedent types and relativised elements)

This first argument does not hold.

The wide range of relativized elements is due to a phenomenon which is a tendancy towards simplification, in particular as manifest among the less proficient users of the language. The word has become an easy "filler" to be used when dealing with all sorts of relativized elements that leave a fuzzy picture in the mind before the day comes where one is perfectly familiar with them; necessarily the role it plays in doing so is the role of the more specific word (or combination of words) . That does not confer to the word a universal nature. There is no reasoning showing that.

Also, the extent to which "that" is used has nothing to do with its role in the clause; it is probably the case that it appeared ulteriorly as a pronoun (the ngrams collected below tend to support this contention (blue curve almost always above zero at the start) and in most cases did not gained more currency than the wh-pronouns, one of the possible reasons explaining that being a want in specificity whereas the greater "uniqueness" of the role of the wh-pronoun gives them that added specificity ("that" is a demonstrative, a pronoun, a conjunction (now called "complementizer")); the fact that there is a modern trend to choose the more specific pronoun (blue curve dominates often at the end) shows that simplification has less appeal (possibly due to an increase in literacy in recent times).

It is said as an argument towards proving the non-pronoun nature of "that" that on top of serving as a "wh"-word, it is also used in constructions where it has nothing to do with those words; but, more generally, "that" as a demonstrative has nothing to do with wh-words ; is that a sufficient reason to say that it must not be a demonstrative? I don't think so. The reasoning adopted in H&P is entirely similar and therefore fallacious.

The fact that there is no grammatical word used to refer to "to you" proves nothing. In the more usual and explicit "It wasn't you (that) I was referring to." the usual relations are clear and they appear neatly in the plain corresponding sentence "Who(m) I was referring to was not you.". The fact that "that" took on a new role in cleft sentences, where wh-words fail to render the needed effect, is a proof of nothing. In "It was the dog to which I gave the water.", "to which" keeps its relative restrictive role, and therefore cannot be a marker of the cleft-construction. It becomes evident that the simplifying use of "that" in this sentence would blur the meaning, although in speech, main stress on "dog" tends to clear up the ambiguity.

the girl who/that spoke first

the book which/that she read

the day when/that she arrived

the reason why/that he resigned

the reason why/that she was

II (Lack of upward percolation)

The second argument fares no better.

As seen in "I", "that" has very probably been introduced or used consistently enough after the more usual pronouns had been established, possibly at first as simplifications that suited less proficient users of the language. This is usage; why should the locutors have introduced different forms of "that" made to correspond to each one of the wh-words? Why this redundany in the first place? If the context was that of simplification and relaxed language, there was certainly no interest in systemization, and the lack of these forms in the language is quite natural; that does not take anything away tfrom the initial nature of pronoun of "that". The other contention, namely, the impossibility of combining "that" with certain prepositions in the way that is possible for the wh-word would reinforce the belief that "that" is not a pronoun, does not hold water either. Here again, we are dealing with usage. "That", let's not forget, is originally essentially "the" (Etymonline); it is clear that several shifts took place from article, to demonstrative, and finally to relative and conjunction. We can speculate as to what caused combinations such as "with that", "to that" and "that's" to have remained inexistant when "that" is a relative; perhaps this is so for the first two because they occur in plenty when "that" is the demonstrative ("to that", "with that"), or because they appeared to be too complicated (if simplification was the main idea, this is even more likely)—and as concerns "that's" see note "9" above in H&P's text (proof that usage is king)—, but it is clear that if they begin to be used nowadays, no principle can keep people from doing that, because there are none; these new combinations will be legitimate new constructions of the language. Therefore this is merely a question of usage.

III (Finiteness)

The question of what verbal forms are usable in a relative clause introduced by "that" yields no more light on the nature of "that" than the preceding argument (II); the reasoning is similar. Why it is not acceptable to say "That's the knife with that to cut the vegetables" is a matter of usage, strictly. The day the user's mind can get to the point of apprehending easily that in this verbal context "with that" is not a PP and that "that" can't have but the role of relative pronoun, there is a new legitimate usage in the language and "that" is certainly a pronoun in it. The fact that both finite and nonfinite verbs are used with the wh-pronouns is inconsequential: it is not unlikely that in most languages this would have been achieved in two steps, first the finite verbs, and later the nonfinite.

IV (Omissibility)

The possible omission of "that" is certainly a matter of simplification, and it has so far been considered to be an informal practice; it is in harmony with the idea that what prompted this word into the rank of the pronouns could have been simplification. There is again nothing inherent in the various possibilities of omission that point to a particular grammatical nature of this word.

---

CRITERION

(CoGEL) The chain relationship is an 'and' relationship, whereas the choice relationship is an 'or' relationship. Thus if two units X and Y occur one after the other in a larger unit, they are in a chain relationship, X + Y. But if X and Y can be substituted for one another in a larger unit, they are in a choice relationship, X/Y. (Substitution here means 'commutability', ie acceptable replacement in terms of the structure of the sentence, not necessarily in terms of meaning.)

Given the sentences

• "They told the only person who was left in the room to leave right away.", and
• "They told the only person that was left in the room to leave right away.",

it can be said that "who" and "that" are commutable; moreover the meaning is the same in both sentences. This is a garantee that "that" is a pronoun.

Similarly, "They told a person whom they knew as being a relative." and "They told a person that they knew as being a relative." show that "that" is the accusative relative pronoun. However, "They told a person whom they knew." and "They told a person that they knew." does not allow to say anything about "that", except that it is either pronoun or complementizer.