• Sandhya Sundaresan

Thoughts on de se: Part I

If I talk about myself or think about myself or dream about myself or see myself in a mirror, am I aware that the person I am talking or thinking or dreaming about or seeing, is actually me? What do the conditions of the world have to be like such that I am (not) capable of correctly identifying myself? And, more interesting for us linguists, what do the conditions of grammar have to be like such that an attitude-holder is aware that a "self-attitude" (an attitude someone has about themselves) is a self-attitude, in a sentence that expresses this attitude? And perhaps even more peculiarly, what are the grammatical conditions under which, in such a sentence, the attitude-holder doesn't have a choice but to be aware that a self-attitude is a self-attitude? The former kind of proposition, which allows false self-identification by an attitude-holder, is compatible with a de se or de re reading; the latter, which disallows false self-identification by an attitude-holder, yields an obligatory de se reading.

For someone who is not a semanticist (even if I do work on issues at the syntax-semantics interface) and also finds deep discussions about possible worlds somewhat scary, I seem to spend a lot of time thinking about these questions. Some of this isn't so surprising: I care a lot about the grammar of reference, and from there, it's a short (even trivial) step to thinking about the grammar of self-reference; I also care a lot about finiteness and clause-structure: and it turns out there is a very deep connection between the grammar of self-reference and that of the clause. To rehash a now-familiar distinction, the (prototypically non-finite) control complement of an attitude verb like "want", as in (1), forces a de se reading: to reframe it in the terminology I've introduced, Maria doesn't have a choice but to know that the (imaginary) person who chugs beer is the same as herself, the person who wants this beer-chugging to happen. In (2), which involves a prototypically finite "that"-clausal complement, false self-identification is possible: Maria is still thinking a thought about herself, but Maria doesn't need to be aware that this thought is about herself:

(1) Maria_i wanted [PRO_i to chug the beer].

(2) Maria_i thought [that she_i would chug that beer].

What interests me is understanding why the properties of the clause seem to affect the possibilities for correct or false self-identification, in precisely this way. Why does the possibility for a de se reading seem to incrementally decrease with the size of the clause (assuming of course that a "that"-CP is larger than a non-finite one), and how do we model this? A lot has been written about precisely this -- but it has been primarily, perhaps even exclusively, by semanticists or philosophers of language. Syntacticians, for the most part (including me, until recently, like I said), tend to give this stuff a wide berth: the general attitude is that this is scary, complex semantic stuff involving modality and possible worlds and way too many lambdas (and, shudder, with superscripts!)-- thus way above the pay-grade of a humble syntactician (or way below the pay-grade of a "serious" syntactician -- take your pick). But is it entirely obvious that these issues lie beyond the pale of syntax? What kinds of grammatical traces would de se have to leave behind to convince even the most hidebound, territorial syntacticians that it, or part of it, might need to be implemented syntactically? And are such effects actually attested? More weakly, is there conclusive evidence that the de se/de re distinction is handled in post-syntax (LF and discourse-pragmatics) and not in syntax?

This is what I want to talk about in my next post (spoiler alert: I don't actually have concrete answers! But I haven't seen serious discussion of whether the de se/de re distinction is part of syntax, literally anywhere. I want to talk about why that is, and discuss a couple suggestive pieces of evidence that may push us to revisit the question).

In this post, I want to talk about something else that interests me about all this, which also doesn't really get discussed as much as it should: the (dubious) relevance of the de se/de re distinction in pop-culture and the "real" world. There is a common notion (I have definitely seen more than one syntactician roll their eyes when discussion of the de se/de re distinction rolls around), that this is not a particularly useful thing to talk about (and that its uses are more purely academic in nature -- maybe even just a fallout of the power and elegance of intensional semantics). This is actually understandable: the discourse contexts under which someone would actually mis-identify themselves are, for the most part, contrived or improbable. These scenarios often involve individuals in various stages of confusion (a spy named Ortcutt who ends up spying on himself, or a philosopher (John Perry) who ends up chasing his own sugar trail in the supermarket); amnesiacs (who have forgotten their true identity, thus can be put in situations mis-identifying themselves), are another hot favorite; or they involve you-had-to-be-there-to-believe-it situations that are so unlikely, they get written about in the media -- like this and this. Testing this kind of thing among native speakers in the field is consequently also a nightmare (trust me, I've tried).

But one thing this quarantine has taught me is that, within the realms of fiction, especially fantasy fiction, these scenarios are neither improbable nor contrived, so maybe this is where we should be looking for examples. The two of us have been listening to a lot of Harry Potter (Stephen Fry's rendering, which is simply brilliant, by the way) and false identity or complex identity scenarios are all over the place in J.K. Rowling's magical world! There are scenarios that test the classic de se/de re distinction and blend seamlessly into the narrative (after all, J.K. Rowling didn't write these in to showcase the de se/de re distinction!). And there is even one recurring theme which tests something that doesn't get discussed, because the conditions that would make it true are impossible in the actual world (but possible in a magical world, like Harry Potter's). This is what I'll call the reverse de se effect.

Some quick definitions to clarify what I mean and to frame the discussion following:

- Classic de se: an attitude-holder has a first-person attitude about herself (she knows the attitude is about herself, so it can be rephrased as a first-person attitude)

- Classic de re: an attitude-holder has a third-person attitude about herself (she doesn't know the attitude is about herself, but thinks it's about someone else)

- Reverse de se: an attitude holder has a first-person attitude about someone else.

Here's a lovely classic de re scenario from Harry Potter that I particularly like. It's from Harry Potter & the Prisoner of Azkaban and it has a cool scene where Harry and Hermione time-travel to the recent past. Stuff goes bad, as it often does, and toward the end of this time-travel adventure, past-Harry is lying facedown on the ground near a pond while the evil dementors descend upon him and get ready to give him the infamous Dementor's Kiss (which, of course, rips apart your soul, in case you didn't know!). As he struggles and prepares for the worst, a Patronus (which...never mind, just look it up if you don't know) in the form of a stag races across and drives the dementors away. Harry looks across the pond and sees someone who looks eerily familiar and similar to himself, who has obviously summoned the Patronus. He mistakenly believes that this person is his father, miraculously come back to life, but it turns out to be Harry himself -- future-Harry from the narrative-now. This is a classic case of mistaken identity (classic de re): Harry sees someone that is actually himself but mistakenly believes it's someone else. And it's not contrived at all, but makes perfect sense within the context of the narrative. Indeed, it couldn't have been any other way, given the surrounding narrative context.

The reverse de se is part of a recurring theme and major plot-driving device. It involves the use of the Pensieve. If you aren't a major Potter nerd (Potterhead?) like me, this is basically a basin that someone can fill with their thoughts and memories. Someone else could then plunge their head into the Pensieve and experience these thoughts. The Pensieve is used quite a bit in Harry Potter and the Half-Blood Prince. In this book, Harry gets special one-on-one lessons from Dumbledore, all of which involve him, together with Dumbledore, plunging his head into the Pensieve to experience someone else's thoughts or memories -- either of Dumbledore himself or of some third, salient individual, e.g. Prof. Slughorn. These Pensieve sessions involve a fascinating interplay of identity relations. When Dumbledore plunges his head into the Pensieve holding his own memory, we have a case of classic de se: Dumbledore revisits his own past experiences and he is furthermore aware that these experiences are his. This is simply a more vivid variant of someone having a nostalgic flashback: given the constraints of the physical world we live in, we can't actually physically transport ourselves back to that time, but we can see aspects of what we lived through in "our mind's eye". Close enough.

But when Harry has his head in the Pensieve and revisits Dumbledore's memories, we get a reverse de se effect: Harry experiences the world from Dumbledore's perspective, not his own. I.e. he experiences Dumbledore's memories not second-hand, e.g. by having Dumbedore report these to him at a later time, but first-hand, from within the mind of the Dumbledore who lived those memories back when they happened: he sees the people Dumbledore saw then, he goes where Dumbledore went. This is, of course, impossible in the real world: you can't go inside someone else's mind (at least not until we develop telepathic communication, and maybe not even then). And if I were to come up with a scenario where someone did do so, you'd call it impossibly contrived, even ridiculous. But within the magical world of Harry Potter and his friends, it blends in seamlessly, and makes eminent sense. It is a major plot-driving device which affects the sequence of events in the final book, shapes relationships, and conditions character-awareness.

What's interesting to me about the Pensieve is that, even though Harry is literally inside Dumbledore's mind, he still remains Harry. He is in the world of Dumbledore's mind, but he is still himself --- observing as Harry, having feelings and opinions about what he sees as Harry, jumping in surprise at an experience that surprises him, or shouting out loud if someone in the memory is in danger. It creates all kinds of weird cognitive dissonance. What is 1st-person and what is 3rd-person and how do we decide?

Of course, you could say that these types of example don't challenge the idea that mistaken identity scenarios are very improbable in "real" life. And it's true I had to use examples from a famous fantasy series to make my point. But mistaken identities play an important, even definitional, role in a more somber aspect of real life: cases of schizophrenia. I realised this after listening to a couple of talks by Anders Holmberg about "self-talk", most recently at an Olinco in Olomouc. A schizophrenic is a schizophrenic because she hears thoughts inside her head that are actually hers but that she attributes to someone else ("voices" in her head). Conversely, those of us who are not schizophrenics are not schizophrenic because we are aware that our self-thoughts are our own and not someone else's.

So issues of de se/de re aren't peripheral or "begat by elegance" (to borrow an expression from David Kaplan when talking about monsters and indexical shift) -- they are fundamental to our cognitive well-being, something we take for granted every day. Stay tuned for Part II...


Homepage of Sandhya Sundaresan.