Special counsel Robert Mueller wants you to know that he thinks President Trump committed obstruction of justice

Bob Mueller finally speaks. The main headline is “Mueller says he did not exonerate Trump!” This isn’t news though—Mueller already says this loudly and clearly in the written report. So does this press conference have anything to add to the written report? Yes and no. No in that, the content of what Mueller says is no different than what’s in the report, and in fact it seems that part of his goal in speaking is to restate the main findings of the written report, and to advertise the report itself. One could imagine an SNL skit reenacting the press conference in which Mueller walks up to the podium and shouts “Go read the report!”, before turning around and disappearing from public life forever.

Still, I think hearing Mueller give a very brief summary does add two things. First, the brevity of the spoken remarks forces Mueller to give the most succinct account of his results and reasoning to date. Having it all there in a few hundred words makes it easier to draw some pragmatic inferences about his unspoken thoughts on the basis of what he actually did say. Second, as a mode of communication, linguists take speech to be primary to writing. Part of the reason for that is that the unique medium of making vocal sounds offers the speaker tools of communication largely absent in writing, especially prosody--intonation and stress. Mueller makes interesting use of these at a key moment. I'll address each of these points in the following from a semanticist’s perspective.

You can read Mueller’s full comments here, and see the video here. You can find the full written report here. It will be helpful to have the relevant 348 words to refer to, so I reprint them just below:

The order appointing me Special Counsel authorized us to investigate actions that could obstruct the investigation. We conducted that investigation and we kept the office of the Acting Attorney General apprised of the progress of our work.

As set forth in our report, after that investigation, if we had confidence that the President clearly did not commit a crime, we would have said that.

We did not, however, make a determination as to whether the President did commit a crime. The introduction to volume two of our report explains that decision.

It explains that under long-standing Department policy, a President cannot be charged with a federal crime while he is in office. That is unconstitutional. Even if the charge is kept under seal and hidden from public view—that too is prohibited.

The Special Counsel’s Office is part of the Department of Justice and, by regulation, it was bound by that Department policy. Charging the President with a crime was therefore not an option we could consider.

The Department’s written opinion explaining the policy against charging a President makes several important points that further informed our handling of the obstruction investigation. Those points are summarized in our report. And I will describe two of them:

First, the opinion explicitly permits the investigation of a sitting President because it is important to preserve evidence while memories are fresh and documents are available. Among other things, that evidence could be used if there were co-conspirators who could now be charged.

And second, the opinion says that the Constitution requires a process other than the criminal justice system to formally accuse a sitting President of wrongdoing.

And beyond Department policy, we were guided by principles of fairness. It would be unfair to potentially accuse somebody of a crime when there can be no court resolution of an actual charge.

So that was the Justice Department policy and those were the principles under which we operated. From them we concluded that we wouldwould not reach a determination – one way or the other – about whether the President committed a crime.

First, inferring what Mueller thinks from what he said: About the crime of obstruction of justice, Mueller said, “If we had had confidence that the President clearly did not commit a crime, we would have said so. We did not, however, make a determination as to whether the president did commit a crime.” This amounts to the following:

(1) No exoneration: It’s not the case that Mueller believes that Trump didn’t commit the crime of obstruction of justice.

Does (1) mean that Mueller thinks that Trump is guilty, as in (2)?

(2) Guilty: Mueller believes that Trump did commit the crime.

In just a moment, I'm going to argue that Mueller is in fact trying to convey (2) without coming out and saying it. This argument will be based entirely on an analysis of Mueller's use of language in context. (A legal scholar might come to the same conclusions as I do on the basis of the obstruction facts that Mueller laid out, in combination with an understanding of whether those facts would ordinarily support an indictment. Such arguments and conclusions are interesting in their own right of course, but are different from what I have to offer here.) But to get there we need to move slowly by first noting that while (1) is consistent with (2), it definitely does not entail (2). In fact (2) entails (1). If Mueller believed Trump were guilty and therefore deserved to be indicted, it follows the Mueller does not exonerate Trump.

But the truth of (1) is consistent with another possibility, one that is mutually inconsistent with (2), namely (3):

(3) Inconclusive: It’s not the case that Mueller believes that Trump didn’t commit the crime AND It’s not the case that Mueller believes that Trump did commit the crime.

While (1) is consistent with (3), it does not entail it. Again, it is in fact the case that (3) entails (1). If Mueller thought that the evidence was inconclusive either way, then of course it follows that Mueller does not exonerate Trump.

So, (1) is consistent with either (2) or (3), and each of (2) and (3) entail (1). But of course (2) and (3) can't be true at the same time. Either Mueller has concluded that Trump is guilty or that the evidence is inconclusive, but not both.

Now, why didn't Mueller exonerate Trump? Is it because he thinks Trump is guilty (2), or because he thinks the evidence is inconclusive (3)? Remember the quote above (1). Mueller said that they “did not make a determination” either way. Does this mean that he thinks the evidence was inconclusive, as in (3)? No. Don’t forget that Mueller stressed that the reason that they didn't make a determination is that they're not allowed to accuse Trump due to Department of Justice (DoJ) policy:

(4) “A president cannot be charged with a federal crime while he is in office. That is unconstitutional... Charging the president with a crime was, therefore, not an option we could consider.”

In other words, Mueller is going out of his way to remind everyone that he couldn't say (2) even if he wanted to. So it would be premature to conclude from the “did not make a determination either way" comment that Mueller is saying (3).

The upshot is, we just can’t tell from (1) alone whether it holds because (2) holds, or because (3) holds. Is there some way that we can figure out which one Mueller believes? Yes. Suppose for a moment that Mueller actually believed that the evidence was inconclusive, as in (3). If that were the case, then why didn’t he say that? He’s certainly allowed to say, “I am neither confident that he didn’t commit the crime, nor am I confident that he did.” It’s reminiscent of the standard GLOMAR response that government officials give all the time: I can neither confirm nor deny that he committed obstruction. Saying (3) would end speculation about whether Mueller actually believes that Trump committed an indictable offense, which would in turn help Americans turn the page on this mess. So why didn’t he do that?

The only reason I can imagine for Mueller not to say (3) is because he doesn’t believe it, in fact, he believes it is false. Again, there's no legal or policy reason why he can't say (3), and there's massive social pressure to say it if he thinks it's true. So he must not think it is true. What that means is that he has an opinion:

(5) Opinion: Mueller believes that Trump committed the crime OR he believes that he didn't. (For those familiar with logic, note that (5) is just the negation of (3), an instance of DeMorgan's laws)

Remember, Mueller told us (1): it's not the case that he believes that Trump didn't commit obstruction. That fact combined with (5) entails (2): Mueller believes Trump committed obstruction.

To recap: what Mueller has said publicly is the only thing he’s allowed to say that both respects DoJ policy and comports with his beliefs, namely that the evidence doesn't clear Trump, and that he’s not allowed to accuse him of obstruction. If he believed the evidence were inconclusive, as in (3), he could have and should have said that—there's no reason for him not to. Since he didn’t, we can infer that he think the evidence is conclusive, that is he has an opinion, as in (5). That combined with (1) means he believes (2), that Trump committed obstruction. The only reason he didn’t “make a determination as to whether the president did commit a crime,” is because of (4), DoJ policy barred him from directly asserting (2).

Second, Mueller’s use of prosody further conveys his unspoken thoughts: I think the above by itself already strongly implies that Mueller thinks Trump’s actions would warrant an accusation of obstruction, if only he weren’t a sitting president. But, as if that weren’t enough, Mueller makes one final remark on the topic of obstruction, one that makes clever use of English to convey his unspoken intent, though it is easy to miss:

(6) “So that was Justice Department policy. Those were the principles under which we operated and from them we concluded that we wouldwould not reach a determination, one way or the other, about whether the president committed a crime.”

At first glance, the repetition of “would” in (6) seems like a verbal tic. Earlier in the press conference, Mueller seemed to stumble over his prepared remarks a few times, so you could be forgiven for thinking that the repetition was accidental on the first listen.

But watch the video again. The repetition is clearly deliberate. Mueller is stressing the word “would” as well as repeating it. Both behaviors are linguistic markers of emphasis in English. Now why would Mueller emphasize the word “would”? In linguistics, we have a notion called “contrastive focus”—when a word or phrase is phonologically stressed so as to contrast it with an alternative word or phrase that is salient in the context. In (6), Mueller is contrasting “would” with “could”. He is saying that the DoJ policy hamstrung him and his team, and forced them to conclude that they would not make a determination, but not that they could not make a determination. In other words, Mueller conveys that he could determine whether Trump committed obstruction, and that he did determine that Trump committed obstruction. He just couldn’t come out and say it.

Real words and Scrabble

The other night during a game of Scrabble with my family, I played "foam" on a triple word score. I had the opportunity to expand it to "refoam", which would have netted me some extra points, though I didn't dare since I could guess that "refoam" wouldn't be in the official Scrabble dictionary, and so it wouldn't, for the purposes of the game, count as a "real" word. 

What makes a word real? If you were to ask people this question out of the blue, I think vanishingly few of them would say that it depends on whether or not the word appears in the official Scrabble dictionary. But many of them would probably say it depends on whether the word appears in some other dictionary, probably the unabridged Oxford English Dictionary, or something like it.

But if you think about it, dictionary writers are constantly adding new words to their dictionaries. Do they invent these new words themselves? No! They find them being used by English speakers. Speakers of English invented them. So were those English speakers using fake words until they were added to the dictionary, at which point they became real words? Of course not! Dictionaries do not come first, speakers of a language and the words they use come first. And dictionaries are not for separating real words from fake words. Dictionaries are just informational tools to help people understand words they have never encountered before. So whether a word is in a dictionary is really a measure of how useful the dictionary writers thought it would be to their readers to have a definition for that word. This set of words—the ones it would be useful to have a definition for—overlaps greatly with the set of words that we might want to call "real", but not perfectly.

So if a dictionary can't tell us definitively whether some string of syllables is a word in a language, how can we tell? Here's a stab: a string of syllables is a real word in a language if it is used by speakers of that language as a word. This is basically the correct answer, but it's also an unsatisfyingly fuzzy one. Suppose I start saying "flarsnip" as a synonym for "plate". Is "flarsnip" now a word with the same meaning as "plate"? Could I play it in Scrabble? Intuitively the answer to these questions is no. "Flarsnip" needs to catch on enough with the speakers of the language to be a word in the language. But which speakers? We all know that different groups of people sometimes use different words to convey the same meaning. For example, the insect often referred to as a "dragonfly" is sometimes called "darner", "darning needle", "snake doctor", "snake feeder", or "mosquito hawk". People playing Scrabble in Arkansas might be inclined to accept different words than those playing in my native New Hampshire. 

Then there are words that many English speakers use but that we've all been told aren't words, such as "ain't" (a contraction of "to be" and "not", but close enough to a word for us) or the dreaded "irregardless". Both appear in any complete dictionary of English, albeit with usage notes warning that they are non-standard and to be avoided in formal or careful speech. Are they real words? After all they are in the dictionary and they are used by real speakers of the language! They seem to belong to a special category of word: real words that the authorities tell us should not be real.

So I didn't play "refoam", and when I mentioned that I had wanted to play it, everyone but me agreed that it isn't a "real" word. I protested by arguing that it's easy to concoct a context in which it might be used. Imagine construction workers whose job it is to spray foam on walls. Of course they would use "foam" as a verb: "Pat is foaming the wall," or "Hey Pat, go foam the wall, would ya?". And then one day Pat's boss Chris says, "Hey Pat, I'm going to need you to refoam that wall over there, it didn't stick the first time." Needless to say, this argument didn't convince my competitors, and in the end "refoam" wasn't in the Scrabble dictionary either, which satisfied them. And I agree with them in one sense: for the purposes of a game like Scrabble, it's best to have an authoritative dictionary that settles whether or not a word can be used in the game in no uncertain terms. For the purposes of Scrabble, "refoam" is not a "real" word.

But is "refoam" a real word? Of course it is. Just google it. It's used regularly in the world of speaker system repair. There are 145,000 hits, and several Youtube videos demonstrating how to "refoam a speaker".

The Scrabble dictionary, it turns out, has a whole special section devoted to "self-explanatory verbs with RE- prefix". There are many unexpected verbs in there—"refall", "refeel", "refloat" (each of which my computer's spellchecker automatically changes into something else, by the way)—but not "refoam". If you had to guess, would "refight" be in there? Have you ever heard anyone say "refight"? Have you ever said it? You'll have to check an official Scrabble dictionary to find out.

It's not obvious how to decide which words are real and which are fake. For the most part, the distinction doesn't seem helpful. If a word is being used in a genuine way by a speaker, and is genuinely understood by a hearer, then it's a real word, even though most speakers of the language might not know what it means. Only when a speaker intends for a word to be fake, like "flarsnip" above or many of the words in Lewis Carroll's Jabberwocky, is a word fake (though a fake word may still be a word, depending on how you look at it, see Barbara Partee's discussion of the adjective "fake"). To make matters worse, once you start looking at languages from many different language families, and especially polysynthetic languages in which an entire sentence can be made of a single word, it quickly becomes difficult say what a word even is. 

Susanne Langer on the existential presupposition of questions

I bought a copy of Susanne Langer's 1942 book "Philosophy In A New Key" for fifty cents at The Word Bookstore in Montreal the other day. I had never heard of her before, but the work looks interesting. She seems to be an early philosopher of mind who was interested in applying philosophical logic to symbolism, art and music. 

McGill's Semantics Research Group is shaping its reading group this summer around the theme of questions. So I thought it was fitting that the first page of Susanne Langer's book makes philosophical hay out of the existential presupposition of questions:

'The way a question is asked limits and disposes the ways in which any answer to it—right or wrong—may be given. If we are asked: "Who made the world?" we may answer: "God made it," "Chance made it," "Love and hate made it," or what you will. We may be right or we may be wrong. But if we reply: "Nobody made it," we will be accused of trying to be cryptic, smart, or "unsympathetic." For in this last instance, we have only seemingly given an answer; in reality we have rejected the question. The questioner feels called upon to repeat his problem. "Then how did the world become as it is?" If now we answer: "It has not 'become' at all," he will be really disturbed. This "answer" clearly repudiates the very framework of his thinking, the orientation of his mind, the basic assumptions he has always entertained as common-sense notions about things in general. Everything has become what it is; everything has a cause; every change must be to some end; the world is a thing, and must have been made by some agency, out of some original stuff, for some reason. These are natural ways of thinking. Such implicit "ways" are not avowed by the average man, but simply followed. He is not conscious of assuming any basic principles. They are what a German would call his "Weltanschauung," his attitude of mind, rather than specific articles of faith. They constitute his outlook; they are deeper than facts he may note or propositions he may moot.
'But, though they are not stated, they find expression in the forms of his questions. A question is really an ambiguous proposition; the answer is its determination. There can be only a certain number of alternatives that will complete its sense. In this way the intellectual treatment of any datum, any experience, any subject, is determined by the nature of our questions, and only carried out in the answers.
'In philosophy this disposition of problems is the most important thing that a school, a movement, or an age contributes. This is the "genius" of a great philosophy; in its light, systems arise and rule and die.'

A card trick

Yesterday evening, a friend of my mother’s regaled us with a card trick. (Fair warning, this post will reveal how the card trick works. But it’s not an especially amazing trick. I don’t think my revealing it will be a great loss to the magic profession.) He put a deck of cards in his back pocket, then asked my fiancée to pick two suits, then to pick her favorite of those, then to pick a run of five cards in the suit, then to pick another entirely different run of five cards, then to pick three of those, then two of those remaining three, then one of those remaining two, which were the two and queen of hearts. She picked the two of hearts, and he said, 'That leaves the queen of hearts as the unnamed card. Pick any number up to 52, and I will produce the queen of hearts from my back pocket in exactly that number of tries.' My fiancée, feeling a bit impish under the influence of Glenlivet, chose 51. (Actually, her first choice was zero, which might be interesting to a linguist studying the meaning of "up to 52", but since he clearly wouldn’t be able to produce the queen of hearts in zero tries, he made her pick again.) So the dilettante magician pulled one card after another out of his back pocket, until impatience set in, then he pulled most of the deck out of his pocket, counting out the cards. At about 40, he exhausted the ones in his hand and went back to his pocket. At 51, he pulled out the queen of hearts. We applauded him. How could he have done it? It might seem obvious to the reader. Sitting at the table last night, the explanation wasn’t immediately apparent. But some reflection revealed the answer.

Got it?

The answer got me thinking about magic tricks: Do they all turn on some kind of logical contradiction? The magician makes you believe a proposition that contradicts another proposition that would explain the trick. Due to the contradiction, you reject the explanatory proposition. But now, given everything you know about the immediate context and the world, the magician should not be able to do what he did. The conclusion of the trick is a physical impossibility. That’s why we call it magic, right? Now, if you don’t believe that magicians are people who can break the physical laws of our universe, and if the propositions you believe lead you to the conclusion that the magician broke the physical laws of our universe, then it must be the case that the proposition that contradicts the explanatory proposition is actually false, and that the explanatory proposition is true. 

Let’s take the card trick above as an example. It would seem that the only way for a person to avoid producing the relevant card until they reached the randomly chosen number would be to know where the relevant card was in the deck before placing the deck in their pocket. But we also know that the participant from the audience chose the relevant card after the magician put the deck in his pocket. Let’s set these two propositions apart.

(1) The magician knew the location of the relevant card before placing the deck in his pocket. (the explanatory proposition)

(2) The relevant card was not chosen until after the magician placed the deck in his pocket. (the magician's bait)

Clearly propositions (1) and (2) cannot simultaneously be true (assuming he didn't memorize the position of every card before placing the deck in his pocket).  So we have to jettison one of them. Everyone in attendance saw the relevant card get named. It happened after the deck was in his pocket. So (2) has to be true. Therefore, (1) cannot be true. The magician was able to find the relevant card in some other way. But how? It's this question that produces wonder in the audience. We've rejected one of the contradiction-creating propositions, but what we're left with is an unexplainable result…unless of course you believe in magic. ESP, say. Some sort of paranormal intuition. The magician was able to know, in an extrasensory way, where the queen of hearts was in the deck.

But no one in our society seems to really believe that card trick performing magicians use ESP. Perhaps we have seen too many card tricks explained, or performed badly by friends. (Psychics are a whole different story. Lots of people in our contemporary society seem happy to believe in the paranormal abilities of psychics. But if some people really do have such powers, why isn't there any scientific documentation of it? Why don’t these psychics use their power for something more than charging a room full of believers forty dollars a head for cryptic messages from dead relatives? Oh, and don't forget to buy a signed copy of the autobiography at the merch table on your way out! No, the feats of psychics must work similarly to those of card trick magicians.) So if ESP is out, then how did my mother's friend find the queen of hearts in his back pocket on the 51st pull?

Maybe we rejected the wrong proposition. Let's reject (2), and leave (1) intact. The relevant card was chosen before the magician put the deck in his pocket, and moreover he knew where in the deck it was. Perhaps the explanation was clear to you from the beginning; after all, my mom's poor friend didn't have much luck with my fiancée. He had to ask her to pick a second run of five cards. Then when she chose the two of hearts, he had to resort to making the unnamed card the chosen one, which is a pretty fishy way of having a participant choose a card! The trick wasn't in magically, via a sixth sense, finding the queen of hearts in his back pocket on the 51st pull. The trick was in convincing the audience that the relevant card was chosen at random by my fiancée. It's easy to find the queen of hearts in your back pocket if you know ahead of time that that is the card you'll be looking for and if you put it on the top of the deck.

The trick seems to be in getting the audience to reject the explanatory proposition, but that proposition can never be completely rejected. The audience rejects (1), but they also know that (1) is the only non-magical explanation of how a human being can pull a specific card out of his back pocket at a specific point in a sequence. So we reject (1) and we're left with magic as the only explanation. Since the audience doesn't believe in magic, they know they've been tricked but they don't know how. Even so, the result is impressive.

I imagine professional magicians could spice this trick up a little more. Before putting the deck in their pocket, they'd shuffle it up a bit. Masters of sleight of hand, they'd be sure to keep the queen of hearts on top or on bottom the whole time. They might let the participant cut the deck, but again, they'd put the queen of hearts on bottom. Or maybe the queen of hearts is already in their pocket, and they just keep track of where it is when they stick the deck in there. In "Intuition Pumps and Other Tools For Thinking", Daniel Dennett writes that card magicians keep a few tricks lined up at all times. They start off by trying to do a hard one, one that the percentages say they are unlikely to pull off. If they realize they can't pull it off, they try for an easier one. If that fails, they move to an even easier one and so on, until one works. If they are forced to their weakest trick, then the result will seem less impressive, but the audience will still be pleased. My mom's friend mentioned that if he performs the trick with a female participant, the relevant card is always the queen of hearts. If it's a male participant, it's always the ace of spades. Chances are better that members of each respective gender will home in on each of these respective cards. If my fiancée had been typical, he might have gotten her to pick the queen of hearts herself and quickly, convincing the audience even more steadfastly that (2) is true.    

To me the interesting question now is, do all magic tricks turn on this kind of contradiction? Is the audience always faced with a set of contradictory propositions, one of which they must reject? And is the magician's job simply to convince the audience to reject the wrong proposition? To reject the proposition that will force them to posit magic in order to explain the result? 

Imperative signatures

When I was little, I somehow learned to interpret the word "love" in letter signatures of the form "Love, X" as imperatives. As in, "I demand that you love me." So recently, when I signed a father's day card, "Love, your oldest son," I mentally added, "but love your youngest son too." Of course, this must be wrong. What most letter-signers mean by the penultimate "love" is probably something synonymous with "with love". It's a sort of sentence adverbial modifying the signature (or indicating something about the signer's sentiments about the signature or the whole letter) like "sincerely". I seem to have developed this misinterpretation on my own as a child. Certainly no one said, "...and the word you write before your name is interpreted as a command." So now I wonder, are there others out there (mis)interpreting signatures in the same way? Is this like some sort of speech act eggcorn that could gather steam in the language?