So, to briefly recap: we’ve established that there’s a thing called construction grammar, which can be used to describe the ways people speak and understand language (see Part 1). We also established that there’s a particular type of construction grammar called Embodied Construction Grammar, which bases its idea of “meaning” on schemas about embodied experiences in the world.
Where does profanity come in?
As it turns out, profanity throws a wrench in some long-held ideas regarding the fundamentals of syntax, which is relevant for any model of language, including construction grammar. Chapter 6 of Ben Bergen’s new book, entitled What the F, discusses several problems that arise when trying to account for the grammatical properties of profanity. Below I’ll briefly summarize one of the issues covered in the chapter. This is only one of the issues raised, and a serious consideration of the problem would require treatment of all these issues.
Sentences Without Subjects
One of the oldest ideas in grammar is that every sentence has a subject. The subject is usually the thing directly involved in the main verb (e.g. the action) of a sentence, and is the grammatical object with which the verb “agrees” (e.g. you ran, I run, he runs; a verb can be “conjugated” in different ways to agree with the subject). Sometimes this subject is explicitly mentioned, such as:
I’m tired of these motherfucking subjects in these motherfucking sentences.
Other times, in imperative sentence (e.g. “commands”), the subject is implicit. Commands are usually directed at someone, and this recipient of the command is almost always the implicit subject of the command:
(You) Go to your room!
Certain profane utterances, however, do not have an apparent subject. Consider the sentence: Fuck you! The “you” in the sentence isn’t the subject – the “you” isn’t performing an action (whereas “you” would be the subject in You fuck), nor does fuck have to agree with the word that follows it (again, consider reversing the words; you can say either fuck you or fuck that guy, but if you invert the sentences, you must say you fuck or that guy fucks) – so if there is a subject, it would have to be implicit, as it is in commands.
But this also doesn’t seem to be the case. If fuck you is a command directed at a hapless listener, then the listener would be the implicit subject of the sentence. That means the unabbreviated form of the sentence would be you fuck you. However, we know that when the subject and “object” (e.g. the word “dog” in “he washed the dog”) refer to the same thing, the object becomes reflexive. Contrast:
He washed him.
He washed himself.
In the first sentence, we understand that “him” does not refer to the subject of the sentences; it’s some other recipient of the washing action. Thus, if you was the subject, the object would be yourself, as in fuck yourself. This is, of course, an acceptable sentence, but it’s notably not the same sentence as fuck you.
The chapter discusses why some other possible implicit subjects (e.g. God) don’t work, and also mentions some other strangeness with profanity, such as sentences like: White wedding, my ass! This, as Bergen points out, doesn’t have a clear subject; what’s more, it doesn’t even seem to have a verb. And, of course, profanity is used all the time in one-word sentences, like Fuck or Shit! While these can be construed as commands (you, fuck! or you, shit!), they’re often not directed at anybody in particular.
For people who thought they had a firm grasp on the structure of language, these revelations might be kind of concerning. Bergen leaves us with a couple of options:
- Some sentences don’t have subjects.
- Some utterances (e.g. epithets) aren’t actually sentences. They’re something else entirely.
Either conclusion is a pretty big deal. Conclusion 1 forces us to reevaluate our definition of what a ‘sentence’ is. That is, previously we had a certain model of how to represent the kinds of information that sentences convey (something involving a subject and an action somehow related to that subject), and now a crucial part of that representation doesn’t seem to be necessary. And if we go with Conclusion 2, we have to postulate a new kind of information structure, which operates under a different set of semantic and syntactic constraints, and which is designed to convey different types of information to the listener.
Subject-less Sentences in Construction Grammar
Let’s briefly explore how these conclusions might be handled in something like Embodied Construction Grammar.
Conclusion 1: Some sentences don’t have subjects
From a syntactic point of view, this is relatively straightforward to program into a system. For example, if we’re dealing with a Phrase Structure Grammar, we can simply posit two alternate definitions of English sentence forms:
S1–> NP + VP
This is an oversimplification of how one would actually want to implement grammatical rules for the two varieties of sentences, but it illustrates the essential point. Furthermore, if we have a grammar that allows us to impose selectional constraints on phrasal constituents, we can refine our definition of ‘S2’; we can postulate, for example, that if the VP is Transitive (it has an object), the Verb must be of a particular form (e.g. “fuck” instead of “fucks”). Ignoring some of the other issues raised above – such as sentences without verbs, like “white wedding, my ass!”, and the contention that purely expletive sentences like “Shit” aren’t obviously verbs or nouns – we can assert that these phrase structure rules capture this new distinction.
The first problem that emerges is that now we have more “primitives” in our grammar rules. The more primitives a scientific model has, the more assumptions it’s making about the phenomena it’s trying to model. This is not ideal.
The second problem is that even if we accept this approximation of a syntactic rule, we’re interested in constructions, not meaningless, context-free rules. So what’s the meaning of a subject-less sentence like “Fuck you”? In ECG, Sentence-constructions all have a meaning of something called an “Event-Descriptor”. Simply put, at a very high-level, a Sentence conveys an information structure containing information about some ‘event’. The event can vary in linguistic complexity from things like “Spot ran” to “Spot quickly tired of reading the monotonous blog post, since he had neither the time nor the ability to read, so he galloped from the room out into the yard beyond, where he frolicked among the dragonflies and lilies, and experienced the joys of olfactory overdrive.” Crucially, both sentences convey information about a profiled participant; this is a fancy linguistic term for what essentially amounts to the subject of a sentence – e.g. the participant in an event that the sentence’s structure is highlighting or “profiling”.
But we’ve already established that “Fuck you” doesn’t have a subject, which means that one of the fundamental roles of the Event-Descriptor schema would be unfilled. This suggests that we ought to use a different meaning representation for “S2” sentences – but if that’s the route we’re going, we’re postulating the existence of two same-named constructions that are conveying very different information structures. Given that construction grammar is pretty concerned with meaning representation – and a construction is generally considered a “construction” because it’s frequently paired with a similar meaning – what exactly would S1 and S2 have in common, other than our blind commitment to the claim that they are both “sentences”? What, in this model of grammar, is a “sentence”?
This forces us to consider a second conclusion…
Conclusion 2: There are other utterances besides sentences
As with above, this conclusion results in the proliferation of constructional primitives. We’re now asserting that there are multiple “types” of utterances – traditional Sentence-constructions, and another category we’ll call Sans-Subject. Sentence-constructions convey information about an event; we still haven’t decided what information Sans-Subject constructions convey.
But this isn’t wholly bad. At least, this is more or less consistent with the tenets of construction grammar. Before, we had two constructions that we were claiming were the same kind of thing, despite having radically different meanings associated with them. Now, we have two constructions that we are claiming are different kinds of things. What remains, then, is to determine the meaning representation for Sans-Subject constructions.
I can really only make broad speculations about how one would represent the meaning of these constructions. Our working assumption is that in general, utterances (either Sentence-constructions or Sans-Subject constructions) are uttered to convey some sort of intent from the speaker to the listener. Sometimes this intent is assertive (asserting some fact about the world), sometimes it is interrogative (requesting information or an action), sometimes it is imperative (commanding that an action be carried out according to some specifications), but always, there is an intent behind the utterance.
In the case of profanity, our definition of intent and information structure needs to broaden from the case of “Spot ran”. When a speaker says something like “Fuck you”, or “White wedding, my ass!”, the information and intent conveyed seems to be above a purely semantic level. That is, the utterance is intended to establish and make known a belief, desire, or emotion within the mental state of the speaker. When someone says “fuck you”, they are establishing the existence of some belief/emotion in the larger discourse context, and the relationship between the two individuals.
That all sounds nice (or maybe it doesn’t), but it still doesn’t address how we’d represent the meaning. This is even more problematic because the intention of many of these utterances is so context-dependent. A construction like “Fuck you” might generally be thought of as an expression of anger or some sort of antipathy towards the “recipient” of the phrase, but closer examination reveals that it’s not that simple:
- A friend plays a harmless prank on you; in response, you – with a reluctant smile – say “oh, fuck you!”
- A friend tells you something that seems less than credible. In response, you shake your head and say, “fuck you…”
- A friend commits a heinous atrocity. You say, “fuck you!”
Both (1) and (2) seem to be quite similar – both are more or less friendly responses to friendly deception – whereas (3) is an expression of pure anger. The point here is just that you can’t necessarily define some one-size-fits-all meaning representation for a phrase like “fuck you!” because the intention of the phrase varies so much across contexts.
Of course, this isn’t unique to “Fuck you!”; it’s true of language in general. It also doesn’t mean we need a whole set of “fuck you!” constructions to fit the different contexts. Rather, my take on the matter is that a construction for “fuck you!” can convey some prototypical representation of meaning – that is, what “fuck you!” typically means – and the interpretation of that construction is a function of contextual factors, like the relationship to the speaker and what you think the speaker thinks about you, and so on.
 Bergen, Ben. What the F: What Swearing Reveals about our Language, our Brains, and Ourselves. 2016. Basic Books.
 There are, as always, grammatical oddities. Two examples in English are the pleonastic pronoun and the existential there. Pleonastic pronouns, also known as “Dummy pronouns”, often serve as the subject of a sentence, but it’s not clear exactly what entity they’re referring to, nor how they’re directly involved with the verb. A classic example is the word “it” in the sentence “it seems like profanity’s a pretty big deal”. The existential there refers to there in sentences like: “there is a duck on the pond”. Regardless, both sentences clearly have grammatical subjects, even if those subjects don’t pick out a precise entity in the world or the scope of sentence’s action.