Non-restrictive modification

↵ Back to class homepage

Imagine a pretty simple sentence like "Who's the guy in the tank top?" Now let's think about how it might be interpreted in the following two contexts:

  1. There are two guys standing nearby; one is wearing a T-shirt and one is wearing a tank top.
  2. You have been waiting outside the library for your friend to meet you and go get lunch with you. While you're waiting, you notice your friend chatting with a hot sexy guy who's wearing a tank top. There is no other guy around.

What I'm particularly interested in is how we interpret the modifier "in the tank top". In context #1, it would probably be used to specify which guy you're asking about: you are asking who the tank-top guy is and you are not asking who the T-shirt guy is (maybe you already know who he is, or maybe you just don't care who he is). This is what we call a restrictive modifier: it restricts which person or thing we are talking about.

On the other hand, in context #2, it's just extra information. It's not used to specify which guy you're talking about, it's just additional description. This is what we call a non-restrictive modifier.

In English, modifiers are often ambiguous between restrictive and non-restrictive interpretations. An exception is certain relative clauses in a certain register of English; in prescriptive high-school grammar sorts of classes you may have been taught that you should use "that" for restrictive relative clauses and "which" for non-restrictive relative clauses. In many styles of English, though, this distinction is no longer observed.

Is it an insult?

I find that many expressions or utterances may seem insulting or not insulting, depending on whether a modifier in the expression is understood as non-restrictive or restrictive. For example:

During the protests in Hong Kong in 2019, when we weren't busy being tear gassed and shot with rubber bullets and pepper balls, a lot of protesters would shout "黑警" ("evil cop") at the cops. This is widely considered a derogatory term for cops; if someone calls a cop a "黑警", this seems to usually get interpreted as meaning something like "you are a cop, and cops are evil". This is clearly a non-restrictive interpretation of the modifier.

To see that this is non-restrictive, we can imagine what this same expression would mean if it were interpreted as having a restrictive modifier instead. A protester yelling at a cop and calling them a "黑警" could be intending to mean something like, "You are a bad cop (as opposed to a good cop, which is a kind of person I have no problem with)." Perhaps some people did mean it that way (this protest movement included people from a lot of political orientations, so there certainly was a mix of people who believe the institution of policing is fundamentally wrong / people who believe acab, and more conservative people who support the police in general but are just upset at the "bad" ones; the question is whether any of the people shouting "黑警" were in the latter group).

We can see a pretty much identical ambiguity in English as well, although in American English the corresponding term seems to be more commonly interpreted as restrictive rather than non-restrictive. That is, if I say "He's a dirty cop", this would often be interpreted as meaning that the person in question is a particular kind of cop (a corrupt one who uses his position to personally enrich himself, like a character out of The Shield), without necessarily suggesting that I find cops in general dirty. It could, however, be interpreted non-restrictively, as something like "He's a cop, and cops are dirty." (I feel that the intonation is slightly different if I say the sentence with this intention; with the restrictive meaning I would put more stress on dirty and with the non-restrictive meaning I would put more stress on cop.) Note that my intuitions here are pretty much the opposite for Cantonese "黑警" (evil cop) and English "dirty cop"; this may be due to context (since my experience with this term in Cantonese is mostly limited to a particular moment in history when the attention was on the actions of the police force as a whole, unlike in the somewhat-corresponding American context where media narratives still tend to focus on the actions of individual "bad apple" cops [despite the efforts of many organizers and activists to change the focus to the institution of policing in general]).

Non-restrictive interpretations and implicatures

Enjoy this classic scene from The Lord of the Rings: The Two Towers:

While this scene might make you nostalgic for a certain artifact of early-aughts Internet culture (at least if you are old like me), what I'm more interested in here is Gollum's utterance at the very end of the scene, "You keep nasty chips!". Let's analyze this a little.

First off, let me cheat a little to simplify our lives some. Gollum's statement is not a declarative statement, it's an imperative: i.e., "keep your nasty chips (because I don't want them)!". But later in this analysis I'm going to want us to look at truth-values a bit, and doing that with imperative sentences is a bit tricky. So, to make things easier, let's just imagine Gollum had instead said "I hate nasty chips!", and we'll analyze that instead. (Unfortunately this is in keeping with much of the tradition in semantics and philosophy of language, which mostly focuses on declarative sentences, despite decades of being criticized for that [Austin, Levinson, and Lycan all raise this criticism of traditional philosophy of language]. Anyway, if you want a fun extra challenge, once you've understood our analysis of this easier sentence then you can try to redo it with the imperative sentence!)

So imagine Gollum says "I hate nasty chips!". As we have seen above, this could be interpreted in a restrictive and non-offensive way (e.g., some chips are yummy and some are nasty, and Gollum hates the nasty kind but is not expressing any judgment about the other kinds), or it could be interpreted in a non-restrictive way that is offensive to chips lovers (e.g., chips are nasty and Gollum hates them).

To me, the non-restrictive and offensive interpretation seems more natural and more likely to emerge, at least in whatever context we are assuming for this example. If Gollum uttered this and then Sam criticized him for insulting the dignity of the noble chip, and then Gollum tried to weasel his way out by saying "Hey, I didn't say I hate all chips, I only hate ones that happen to also be nasty!", that would probably sound pretty disingenuous (which, come to think of it, would be pretty in character for Gollum!). So, the restrictive interpretation does not seem like the intended or most natural interpretation in this example.

Let's use some of our formal pragmatics concepts to examine these different interpretations.

I think we could say that the restrictive interpretation involves a conversational implicature (specifically, a quantity implicature), and non-restrictive interpretations involve presuppositions or conventional implicatures. If I say "I hate nasty chips" and mean to convey that I only hate nasty chips and not yummy ones, this seems to be based on interpreting "nasty chips" through what semanticists call "intersective modification" or "subsective modification" (e.g., nasty chips is the intersection of the set of things that are nasty and the set of things that are chips), and then ruling out all the other kinds of chips via a conversational implicature. This implicature is comparable to the ad-hoc implicatures we saw in the "What's a stronger alternative?" module, where e.g. "I have lived in the USA and UAE" can implicate that I haven't lived anywhere else. (But it's not so clear that the example with which we began this module, "Who's the guy in the tank top?", can also be explained in this way, since the restrictive interpretation also seems kind of related to a presupposition associated with the definite noun phrase the guy in the tank top—specifically, a definite expression like this presupposes that there is one unique "guy in a tank top" identifiable in this context.)

On the other hand, the non-restrictive interpretation seems to involve a conventional implicature or presupposition. We could paraphrase "I hate nasty chips" as "I hate chips, which are nasty". The grammatical structure "X, which is Y" has traditionally been analyzed as conventionally implicating, without literally (truth-conditionally) entailing, that X is Y. (The suggested in-class activity for this module will consider other possible ways of explaining this interpretation.)

So we've got two interpretations of "I hate nasty chips." The non-restrictive one suggests that all chips are nasty, and does this via what looks like a conventional implicature. The restrictive one is just a statement about chips that happen to be nasty (and, furthermore, an additional conversational implicature can be made to suggest that there are other chips that are not nasty). The question is, how does a listener choose which way to interpret it? And why does a listener like me have the intuition, mentioned above, that the non-restrictive interpretation is more natural and more likely to arise here?

I would like to suggest that this decision is also pragmatic. Specifically, it seems to me like it uses the maxim of relation. If I just wanted to say that I hate nasty things, I wouldn't have needed to specify "chips". (Indeed, under a restrictive interpretation, "I hate nasty chips" would seem to not only elicit a quantity implicature that I don't hate non-nasty chips, but also a quantity implicature that I don't hate nasty non-chips—e.g., I don't hate nasty fries, I don't hate nasty crisps, I don't hate nasty biscuits, etc. This latter implicature does not seem to arise, though.) I could have just said that I hate nasty things.

Thus, if I say "I hate nasty chips" and the listener and I can mutually assume that I am observing the conversational maxim of relation, then they must assume that I had a reason to point out chips specifically. Presumably that reason is because there is some connection between chips and nastiness.

Or perhaps a better explanation, based on the maxim of quantity, is available. We saw above that the restrictive interpretation also involves a quantity implicature (specifically, an ad-hoc implicature), but the quantity implicature I have in mind here is an entirely different one, not based on "alternatives" at all. Recall that the maxim of quantity means that we shouldn't say too much or too little, relative to what's needed for the conversation at hand. In this case, using restrictive modification to specify that I hate nasty chips (as opposed to yummy chips) is probably unnecessary information to express. This is because it's already obvious from the literal meaning: of course I don't like nasty things, because they're nasty! So, if I say "I hate nasty chips" the restrictive interpretation violates the maxim of quantity: I'm saying something that doesn't need to be said because it's already obvious. So, a listener who hears that will likely search for another interpretation, one that actually is appropriately informative. The non-restrictive interpretation satisfies that need, since it actually communicates something new (the conventional implicature that I think chips are nasty—and perhaps also a suggestion that this is the reason I hate them). This explanation based on the maxim of quantity would also explain why I feel like a non-restrictive interpretation is less likely to arise (relative to a restrictive interpretation) for an utterance like "I hate salty chips", where the fact that I hate salty things is not already obvious from the literal meaning of the word.

Anyway, you should take this all with a grain of salt, because I don't have any references for it; I haven't read any literature about the semantics and pragmatics of non-restrictive modification, so I'm not sure if there are better explanations available or if someone has already discussed and disproved the explanations I'm suggesting here. It's just an interesting idea for you to think about.

Video summary

In-class activities

As discussed in the module, we might paraphrase the non-restrictive interpretation of "I hate nasty chips" in one of the following ways:

  1. I hate chips, which are nasty.
  2. I hate chips, and chips are nasty.

In the first paraphrase, "Chips are nasty" is a conventional implicature (or perhaps presupposition) of the utterance, whereas in the second paraphrase, it's an entailment (part of the literal, truth-conditional meaning). Have students figure out which is a better paraphrase of the original utterance (assuming the original utterance is interpreted non-restrictively). Hint: they can review the "Types of implicatures and their diagnostics" module to see how to test whether something is an entailment or a conventional implicature. (Further hint: if I say "I hate nasty chips" when chips are actually not nasty, would you consider the sentence to be false or just weird?)

If they figure this out and determine that #1 is the better paraphrase (i.e., that the "Chips are nasty" interpretation is not an entailment), then have them try to figure out whether it's a conventional implicature or a presupposition. Hint: recall that, as we saw in the "Presupposition" module, presuppositions and conventional implicatures have a lot in common, but it might be possible to distinguish them based on cancellability. On the other hand, presuppositions and conventional implicatures also might differ in terms of given-ness (presuppositions tend to be about information that is already consistent with the common ground, or they can be used to "slip" some information into the common ground while acting like it's uncontroversial); in this case, "nastiness" is kind of a new topic to the conversation but Gollum seems to be using his utterance to add the idea "Chips are nasty" into the common ground without explicitly saying it. (So, by the cancellability test, "Chips are nasty" seems more like a conventional implicature here, but by the given-ness test it perhaps seems more like a presupposition?).


⟵ Generics
Metaphor ⟶

by Stephen Politzer-Ahles. Last modified on 2021-11-13. CC-BY-4.0.