Remembering Tony Kroch...


If you're not a linguist, you need to know a bit of background to understand this story. If you're a linguist, skip ahead to the paragraph after (4). In syntax, we distinguish between sentences that are grammatical and ones that are ungrammatical. Normally, grammatical sentences sound fine (or, as we say in linguistics, acceptable), whereas ungrammatical sentences sound awful (or unacceptable).

(1) a. Grammatical and acceptable: The cat chased a mouse.
b. Ungrammatical and unacceptable: Cat the chased mouse a.

As Chomsky famously (among linguists) pointed out, there are sentences that are grammatical in the sense that they are produced by the same rules that ordinarily give rise to acceptable structures, but that nevertheless are unacceptable. Let's work up to them. (2) shows that we can modify nouns with relative clauses.

(2) a. the mouse that the cat chased
b. the cat that the dog scared

Now let's take a sentence like (3a) and replace the simple noun phrase "the mouse" by the more complex one in (2a).

(3) a. The mouse escaped.
b. The mouse that the cat chased escaped.

So far, so good. Let's do it again by replacing "the cat" in (3b) by (2b).

(4)   The mouse that the cat that the dog scared chased escaped.

(4) is a sentence that's grammatical but unacceptable. So that's a possible combination of properties. Ok, and now what about the missing logical combination - sentences that are ungrammatical but acceptable? They are generally not mentioned in linguistics. We are now in a position to enter the story.

In the fall of 1983, when I started grad school, Tony was off on a Fulbright in Brazil and coming back in the spring of 1984. In the introductory course to syntax, the instructor replacing him told us that Tony believed that there were sentences that were acceptable, but might be ungrammatical. But she couldn't give an example. From this, I concluded that this man Tony must not have a good grasp of the distinction between (un)grammatical and (un)acceptable, which is a well-known failing among ordinary mortals. I didn't go so far as to wonder how he had managed to get hired at Penn, but I figured that perhaps I could help straighten him out when he got back from Brazil.

In due course, Tony came back and taught the intermediate syntax class. It was a very lively class with only three people in it - Tony, me, and one other student. Tony would sit on the table in front of the blackboard in the old classroom in Williams Hall 616, swinging his legs, and hold forth on the topic du jour. It was evident to the merest child that he prepared (for lack of a better term) for class on his 20-minute commute down the Schuylkill Expressway or perhaps even on the elevator ride up to the department by saying something along the lines of "Well, today, I'll tell them about X”. One day X turned out to be the type of sentence that the substitute instructor hadn't been able to remember. The sentence was from a sociolinguistic interview in Bill Labov's collection.

"That's the only thing they do is fight."


Cherries in the wine

At the old Lebus bar on the corner of 34th and Sansom, Tony introduced me to some wine or other. "Can you taste the cherries?" How tempting to sound sophisticated and say, "Of course". In fact, I had no idea at all what he was talking about. So "No." "Oh", he said, "wait a few minutes for the wine to warm up and the aroma to appear". (He apparently didn't trust me to swirl.) And after I waited a bit and took another sip, there they were - the cherries, just like he said!

Taxonomic mind

In connection with something or other - I forget what, but I remember the room, the old T.A. office in Williams Hall 402 - Tony remarked, "Well, of course, you have a taxonomic mind." Oh! I knew enough about the social structure of the field of generative linguistics to recognize I had just been dealt a mortal insult. Oh! In the olden days there would have been a duel. I fumed for the rest of the day. I complained to him the next day. He apologized. I clearly remember the look on his fact, from which it was completely evident that he was truly sorry. "I didn't mean it as an insult. It's just a statement of fact."

What does a thesis advisor do?

In taking me on as his student, Tony wanted to make sure that I understood his part in the enterprise. "As your thesis advisor, I don't do the work for you. I'm like your coach.” Once again, I had no idea what he was talking about. But it helped me understand what coaches do. Like, say you play basketball. Then the coach is basically your thesis advisor.

Devil's advocate

Tony had a habit of arguing forcefully for one position one day, and then a week or two later arguing equally forcefully for the opposite position. He didn't seem to notice! If you helpfully pointed out what he was doing, it didn't help - he didn't remember, and he didn't even care that he didn't remember!

Students (and perhaps others) tended to find this quite unnerving and even upsetting. For some reason, it never bothered me - rather oddly, given that inconsistency ranks up there for Germans as a sin. It just seemed to me that he was inviting us to use our own judgments.


The way Tony's mind worked was through dialogue. He readily acknowledged that he needed conversation partners on the progress, not perfection road to Truth. Maybe that explains the devil's advocate habit. In the absence of a suitable debate partner, he took on the role himself.

He recommended taking an opponent's argument more seriously than they did themselves. He didn't put this forward as his own idea. I seem to remember that he attributed it to Marx. But in fact it may be in Engels.

For example: Prescriptive grammarians object to double negation as illogical by analogy to the laws of arithmetic. The negative of a negative number is the corresponding positive number. Oh, ok. So triple negation and odd-numbered negation is fine then, right?

He was not shy about using this strategy with his nearest and dearest, as demonstrated in the touché conversation.

Writer's block

Tony helped me a lot with writer's block! He introduced me to the distinction between composing and editing. First get it all out, then edit.

I was having trouble writing up my thesis. Tony said, "Just write up the chapters and give them to me one by one, and I'll put them in this drawer. Don't worry - I won't read them." It worked great. I don't think he ever did. (We did talk about the chapters, though, before I wrote them up!)

Flinching before the lash

Don't. (This was not Tony's own idea; it was what he was taught in the Progressive Labor Party.)

The rule says nothing about flinching at the lash - while the slaveowner is beating you. But don't flinch before - at being shown the lash.

As applied to writer's block, this means you need to state your position clearly without ruining the exposition by interrupting it to ward off imagined responses by adherents of the opposing position. In due course, you will deal with them, treating their position more seriously than they do themselves.

Conceivably this rule explains Tony's preference for Don Giovanni over The Marriage of Figaro, which I was never able to understand. Don Giovanni absolutely refuses to flinch before the lash in the Commendatore Scene (libretto with English translation).

Chasm of centuries

My father was Greek, so when I was growing up, I heard many stories about the Greeks, mostly the ancient ones, and their wondrous sayings and doings. One day, after one of these conversations, I was stunned by the thought that here we were in the (then) 20th century, yet that our minds, through language, were able to make contact with the minds of other people across an intervening chasm of 25 centuries. As a child, I had no words for the experience Today, I would describe it as awe - an Old Testament-type experience in an intellectual flavor. Many years later, I recounted this memory to Tony. The beacon in Tony's mind was always on the lookout for "same or different". How else to explain his instant response, ignoring the red herring of the 25 centuries: "Beatrice, what do you think is going on as we are talking right now?"

Before you ask...

"Before you ask 'Is it true?', you have to ask, 'Could it possibly be true?'" In Tony's judgment, many hypotheses in linguistics fell into the non-starter category.


Tony was telling me how important it was for me to work in my discipline. Mishearing the preposition, I agreed - somewhat sheepishly, but whole-heartedly - that it was very important for me to work on my discipline. Well, as it turned out, Tony provided me with opportunities for both.

Stop defending yourself

If I did things that turned out to be suboptimal, shall we say, I would start explaining myself. Tony had no patience with this, "Stop defending yourself!" Well, how else do you expect me to save face? Tony was simply not interested in saving face. He was interested in getting on with the work. Later, when people would defend their suboptimal annotation decisions to me, I would hear myself saying "Stop defending yourself". For the same reason.

Difficult to deal with

Someone complained that Tony was difficult to deal with. George Cardona was mystified. "Tony? Difficult?"

Same or different?

Tony was fond of saying that there were two results in science. Either two phenomena looked the same on the surface, but were in fact different (like birds and bats, or raising and control in syntax). Or two phenomena looked different on the surface, but were in fact the same (like apples falling on your head and planetary orbits, or wh- movement and topicalization in syntax).

More recently, Alex Kalomoiros tells me, he would say the same thing differently: There are only two numbers: 0 and 1.

Unsolved problems

Tony was fond of saying that you can tell if a field is a science by whether the people working in it agree that the field has unsolved problems.

In linguistics, or at least syntax, he said, people had a solution to every problem and did not the acknowledge the existence of unsolved problems, and so therefore the field wasn't a science.

Progress, not perfection

A corollary of the dictum on unsolved problems was Tony's insistence that in science, we need to be content with approximate solutions.

Matters of notation

The situation was often even more primitive than in Unsolved problems. People thought they had a disagreement, but their alternative solutions were actually notational variants, and they argued about which of the notational variants was the correct one without understanding their foolishness.

Tony was dismayed by notation-mongering. What an annoying reminder of the pre-scientific character of the field.

The following passage of Richard Feynman's cheered him up again (from "The Character of Physical Law", p. 18):

The next question was - what makes the planets go around the sun? At the time of Kepler some people answered this problem by saying that there were angels behind them beating their wings and pushing the planets around an orbit. As you will see, the answer is not very far from the truth. The only difference is that the angels sit in a different direction and their wings push inwards.

Mid-level generalizations

Another aphorism of Tony's was that it is the mid-level generalizations in syntax that were likely to survive. So that was the level of generalization to aim for.

What the hell does that mean? Who knows. For starters, presumably, not confusing same-but-looks-different with different-but-looks-same.

Two favorite expressions

Things were not just clear. They were clear as a bell.

If you got done with one thing, it was time for what? "Next thing".

It occurred to me that...

Tony never (perhaps I'm exaggerating) said, "I thought this" or "I thought that". No. Rather, "It occurred to me that ..."


According to Tony, the proper attitude towards Chomsky was Zen indifference.

At a conference at Princeton in Chomsky's honor in 1986, Chomsky gave a talk. In the question period, Tony made some point or other concerning Tree-Adjoining Grammar and the importance of mathematical formalisms. Chomsky's response was quite dismissive, apparently in the usual way. I was so upset at how Chomsky was wielding his microphone against a former student (who had written a Garland-worthy dissertation on scope ambiguities, for crying out loud!) that I left the room. Perhaps I hadn't heard yet about the proper Zen attitude. Tony, on the other hand, both right afterwards and whenever he later recounted the exchange, was pleased as punch because he had done a good job stating his point on front of the right audience.


After Tony got tenure, he began to exhibit symptoms of mid-life crisis. The conventional character of the attack is almost impossible to overstate. First, he wanted a motorcycle. This, Martha forbade absolutely. A sportscar, she allowed. It was a Mazda RX-7. Red. Tomy zoomed around in it for about a year. Eventually, it was stolen from outside his house, the one on Chew Avenue, before they moved to the grander one at Gorgas and Crittenden, elevated off the street, with a proper entrance hall. Perhaps Martha, with her connections in the neighborhood, had something to do with the theft?


About a year after the fall of the Berlin Wall, Tony and I were at some conference in Europe, and he took advantage of the opportunity to go to Leipzig in connection with a claim (eventually successful) for compensation of property of his mother's family that had been expropriated by the Nazis and never restored by the East German government. I was along for the ride to translate. We found the right office, but the office hours for his type of business were once a week on Tuesday. It wasn't Tuesday. Tony explained to the East German bureaucrat that he had come all the way from America and had only that day in Leipzig. The East German bureaucrat regretted there was nothing to be done; Tony would have to return the following Tuesday. Tony repeated what he had just said. The East German bureaucrat was not used to such a response. "Ja, wenn Sie weiter so hartnäckig sind, ..." (Well, if you continue in this persistent vein...)

Yes? Then what will happen?

What often happened. Tony got what he wanted.


I had just given up my job at Northwestern and was done with linguistics. Sabine Iatridou was leaving Penn for MIT, and the department needed a replacement to teach syntax for a year. Over coffee at a windowside table at the old Barnes and Noble at the top of the hill in Chestnut Hill, Tony was trying to convince me to take the position. Had he not been listening? Had I not made it clear as a bell that I was done with linguistics? Incautiously, I mentioned that I was going to become a Buddhist and how important equanimity was to Buddhists. Bad move! "If you're serious about equanimity, it shouldn't matter to you whether you take the position or not, right? Or for that matter, whether you continue to work in linguistics. Right?"


Where are you?

In the course of the "touché" conversation, we somehow got to discussing the nature of the 'I'. Tony asked: "So, as we're talking now, where is your 'I'?" I went inside my head and looked around and said: "Right now, as we’re talking, there is no 'I' in my head. It's bright inside, but it's empty. All of the 'I' is streaming out of my eyes as I'm paying attention to what you're saying." "Exactly," he agreed.

What did he mean? Maybe the streaming-out-the-eyes business happened to him a lot. Who knows?

Time management

With regard to administrative tasks, Tony generally had one (or maybe two) things that he wanted to get done in a day. The rest of the time, he would patrol the department to see what was up and to try out his ideas on people.

Or play computer solitaire, which never ceased to shock me.

Wrong scale

I was trying to figure out how to pitch the undergraduate introduction to syntax. Some students found the course difficult. Should I make it easier? But then the whole point of taking the class would be lost. Should I make it harder? "Wrong question," said Tony. "Make the course more serious."

Tony Kroch, my hero

In the late 1980s or early 1990s, Tony taught for a semester at Harvard. In the course of that semester, Richard Lewontin invited him to present his work on grammar competition to his (Lewontin's) lab because of the mathematical similarity between grammar competition and biological evolution. Several years later, Lewontin gave a talk at Penn that Tony attended. About an hour after the talk was over, Lewontin was chatting outside the building where the talk had taken place with a group of senior Penn administrators, and Tony and I happened to walk by. Spotting Tony from half a block away, Lewontin hailed him with a joyous cry: "Tony Kroch! My hero!" The Penn administrators were puzzled. Some of them knew Tony. But hero? More like gadfly.

Known issues

It was time to release the Penn Parsed Corpus of Early Modern English. I was playing the part of Casaubon in George Eliot's Middlemarch, always on the verge of publishing some allegedly ground-breaking scholarly result, but too concerned about what people will think to actually do it. In the novel, Eliot eventually kills him off, thus (among other effects) preserving him from the wretched fate of having the mediocre and outdated character of his research revealed to all and sundry. This convenient option, however, I was unwilling to follow in the case at hand.

"What exactly is the problem?" Tony wanted to know. Well, the corpus contained many errors, not to mention the inconsistencies, and so on, and so forth. "Oh, let's just do what they do in software development. Just include a section called 'Known issues' in the documentation. You can work on fixing them after the release."

So I made a pretty perfect list of all the imperfections; it wasn't hard at all. In the course of putting together the list, I fixed some of the known issues, and they never really made it onto the list. But the list is still there. And there are still errors in that corpus and the others that followed it, can you believe it? But the corpora have turned out to be useful in their imperfect form, and each release is better than the last one, and perhaps eventually, the known issues will all be resolved. At which point, I won't delete the section heading, but I will comment it out (until it's needed again).

Moore, I said, have you then apples in that basket?

Tony had Jonathan Miller’s skit on Bertrand Russell by heart, accent and all. At one point, Miller visited Penn, and the Penn administration gave people the opportunity to meet Miller over lunch. To Tony's bafflement, he was the only person on the entire campus to respond to the invitation. But he didn't complain. He was very pleased to have Jonathan Miller all to himself.


At one point, Tony launched a campaigns to get people to distinguish singular "biscotto" from plural "biscotti". It went on for months, but cannot be counted as one of his successes.

Impossible, but useful

Tony and Ann Taylor were coming up with annotation guidelines for the parsed corpus of Middle English that they were building. Middle English has rampant variation between OV and VO phrase structure. Where to put the traces of wh- movement in the annotation? Before the verb? After the verb?

Tony came up with an idea. "We'll just put the trace at the beginning of the sentence, before the subject, where it can't actually ever really go, but where we can search for it." So clever!

That was grace

I was describing an experience of attaining a difficult insight to Tony as having the quality less of an act of perception than that of a purely receptive revelation. "Oh," said Tony, "that was grace." It was, too. But where did he get this way of talking? From the Quakers?

Living in a novel

One day, I mentioned to Tony that more and more, I seemed to be living the life of some character in a novel - an Alexander McCall Smith novel, in my case. "Oh sure," said Tony. I was so flabbergasted that he knew what I was talking about that I never asked him who the author of his novel was. Jane Austen, his favorite?

What's the rule?

Two trees had just been planted in front of the house where I live, and I was helping them make it through their first summer with Treegators. But how often should I fill the Treegators? Once a day? Once a week? What was the rule? So German!

Tony took pity on me (and the trees) and effortlessly invented a tree-tending strategy for me: "Just observe the leaves. When they start drooping, the trees need water." Just as Yogi Berra said.


At one point, I took it upon myself to learn the first stanza of the Marseillaise by heart, and Tony and I were amusing ourselves by belting it out together. In the line "Contre nous de la tyrannie l'étandard sanglant est levé" (Against us the bloody flag of tyranny is raised), Tony pronounces the "t" is "sanglant". What? Liaison across the subject and the predicate? That can't be right!

Hmm, let's listen to how Mireille Mathieu sings it, shall we? (liaison at 0:14-0:25)

In respects other than this liaison, by the way, Tony's French was not standard. He learned French for a month in Paris at the Alliance Française. But he became fluent in Senegal, where he collected myths among the Bassari with an eye towards studying their narrative structure. There he acquired clitic doubling, a trait of African French (or was it subject doubling?).


Speaking of the Marseillaise, Tony once waxed lyrical to some Québecois colleagues about the Marseillaise scene from Casablanca (liaison at 0:31-0:42, but harder to hear than in Mireille Mathieu's rendition). Such a stirring scene! His remarks fell completely flat. As Tony realized later, why had he expected them to find the scene stirring? Just because they spoke French? But their ancestors had left France a century or even two before the Revolution.

It's all Trump's fault

Tony gave a presentation culminating in a graph that elicited a gasp of appreciation from the audience. In due course, the organizer of the conference invited Tony to submit a written version of the talk in a special journal issue. Tony wasn't satisfied with the publisher's contract. The organizer of the conference negotiated a contract to Tony's specifications just for him. Every few months, she would remind Tony of the paper he owed her. "I'll have it for you when classes are over." The other authors grew restive. Careers and promotions were on the line. "Yes, I understand, I'll have it for you in a week." Finally, run to ground, Tony offered a final excuse, "The political situation..." On hearing which, wisely, the organizer took mercy on everyone involved and published the issue without Tony's paper.

Your mother knows what she's doing

My mother wanted to give my sisters and me some money. We weren't having it. Didn't she understand the unfavorable tax consequences, blah, blah, blah? It's not like we knew what we were talking about; we were just making stuff up to make her look stupid.

I mentioned my mother's misguided ideas about money to Tony, and he said, "What are you talking about? Take the money. Your mother knows what she's doing." My mother? Know what she is doing? That can't be right!

"Talk to a lawyer," was Tony's advice. I talked to a lawyer. "Oh, absolutely, you should take the money," advised the lawyer. I called my mother to tell her she was right. My mother was not used to being told that she was right. As a result - can you believe it? - our relationship improved quite a bit.

Mai più

The second-to-last time I saw Tony was on his 75th birthday, about three weeks before his death. He blew out the candle on the birthday apple pie and ate quite a large slice. Eventually, the non-linguists left the room.

"I'm sorry," he apologized in a matter-of-fact way, "I'm no longer able to contribute to the work."

Oh, Tony.


"But Tony, no-one is expecting that of you. There is no need for you to worry. You set things in motion, and we'll continue to keep them going." Hmm, he probably thinks I'm saying that just to make him feel better. "Listen." About a week earlier, Joel Wallenberg had published a paper with some collaborators that for the first time derives - derives! - a Constant Rate Effect (in the change from OV to VO in the history of English) from considerations of information theory, a topic near and dear to Tony's heart during his last years - specifically, from the optimization of information density. On hearing the result, a smile appeared on Tony's face, as quick as a reflex, as impersonal as the sun coming out from behind clouds. What joy - a mathematical result in linguistics!

It might help

Shortly after the mathematically induced smile, Tony went to lie down. A few days earlier, the doctors had told him that the pills that had been successful against the cancer for many months had stopped working, as they had told him from the very begining was likely to happen eventually. As a result, he was set to enroll in hospice at home the next day.

Once he had settled down comfortably, he called Joyce, his caregiver, and asked her to find the pills so he could take one. I was scandalized! By definition, hospice means giving up on curing the disease and focusing on palliative care. Here he was, working around the rules as usual. Ever the pragmatic optimist, Tony swallowed his pill and smiled another smile - equally impersonal, but this one full of sly cunning. "Who knows? It might help."


After Tony's initial cancer diagnosis in September 2019, we talked on the phone in the way we were used to. At first every day, and then every other day, and then eventually there came a week when I didn't hear from him all week. Step by step, it seemed to me, Tony was retreating into a primeval forest, the forest of memory, shape-shifting into some wild creature, perhaps the fox that lived in his backyard that he caught on camera one morning a few years ago - the fox and the tree behind him and the lawn all slightly out of focus, the fox's front legs growing out of the earth, autochthonous pillars, the rest of him flowing off like a stream, the great ears listening, listening, the inscrutable muzzle, and finally, under slightly furrowed brows, the eyes, intent and alight with the authority of existence.

Phantom Tonys

I'm finishing a quick email message while waiting for Joel Wallenberg to join a Zoom meeting. The email window is covering up the Zoom window. "Beatrice?" "Hold on, Tony, I'll be with you in a second." I click on the Zoom window. Wait, I'm confused. It's Joel?

I'm chatting in my office with Alex Kalomoiros (properly distanced and all). Oh, look who's coming out of the neighboring office. "Oh, Tony! Come on in." Oops, it's Spencer Caplan, who's just come down the hall and not out of Tony's office at all, which in any event hasn't been Tony's for a year now.