Neville Morley's Blog, page 7

Sheep

The indispensable starting-point of my project on GenAI and the assessment of historical skills last year was student consultation: via surveys, focus groups and some really excellent volunteers on the steering group. Key principle: how do we respond to this development if we have no idea what students are actually doing, why they’re doing it, how they feel about what’s happening, how they understand assessment and so forth? This immediately yielded the vital information – necessitating a hasty refocusing of the rest of the project – that the vast majority of history students were clear-eyed about some of GenAI’s major limitations and certainly not using it to write their assessments (so, panicky reversion to old-fashioned sit-down exams makes little sense). We need to address the actually-existing problems (arguably, more serious ones) rather than the problems that we instinctively imagined. What, incidentally, does it say about our attitude towards students that so many academics immediately assumed that they would all seize the opportunity to cheat?

Now, obviously the project was not grounded solely in student views and experiences; equally indispensable was a clear sense of what studying history at university is supposed to involve and achieve, with a basic assumption that this should not change in essentials. There’s always a gap between what students can do and what we want them to do, and often a gap between what they think they should be doing and what we want them to do; bridging that gap is, glibly, what the teaching stuff is all about. And it’s in that gap that GenAI has a potentially disruptive effect: if it creates the appearance of learning without the reality, or substitutes for learning in ways that aren’t recognised until too late. Or, optimistically, if it offers a means of doing mundane stuff more effectively/efficiently (GenAI as spelling and grammar checker?), saving time for more serious learning, or even offers new but entirely valid enhancements to the learning process. So of course the third plank of the project (and certainly the shakiest) was my own exploration of its capabilities, trying to get a sense of how far e.g. an AI summary of an article might be useful or reliable or GenAI feedback might be genuinely informed and constructive – and, if not, how should we respond to the fact that this is what many students are using it for?

I’ve been prompted to go over this ground again by reading a piece from HEPI discussing a report from QS: ‘Universities, Students and the Generative AI Imperative’ (link). What struck me – besides the fact that they continue to treat ‘students’ as an undifferentiated mass (QS gathered data on discipline – striking preponderance of business and management students – but doesn’t break down any of the results), whereas I am pretty sure that there are significant internal differences – is the fact that they surveyed academics as well. On the one hand, that makes sense; acknowledgement of the basic principle that academia develops its own guidelines and norms for research and teaching through discussion and debate within representative organisations rather than through top-down imposition (how far this works in practice is a different matter, but it’s certainly the idea); so, this can be seen as a contribution to ongoing discussion, and indeed my own project included surveys and focus groups with academics as well (to which no one responded, so a crucial weakness of the project is how far it rests in the end on my personal views about learning and teaching and the capabilities of GenAI in this area).

But. At any rate the way in which the results are presented makes it seem as if the two groups are being treated identically in structural terms: their experiences, understanding, practices, views and anxieties around GenAI are surveyed, revealing a landscape of confusion, inconsistency and fear, which is the problem to be addressed by the wise experts – NOT that the academic ‘conversation’ is (part of) the process of developing responses to the student ‘conversation’. Both groups, in this world-view, need to be instructed and managed, according to pre-existing assumptions – the fact that the title includes ‘The Generative AI Imperative’ – i.e. you have no option but to respond/submit to this change – is all too instructive.

This issue is most obvious (as Margot Finn – @eicathomefinn.bsky.social – pointed out on Bluesky) when it comes to the issue of the ‘ethical’ use of GenAI, where it is assumed (1) that such a thing is obviously possible, and (2) that the question is quite a narrow one, around academic integrity and open declarations of whether or not one has used such tools, not needing to take into account things like environmental impact. ‘Ethical’ is just the thing we must reassure you that we have taken into account so you don’t have to worry about it. Would we be making these recommendations if we thought they were unethical? Of course not!

At this point, I would make a connection to a good piece by Kevin Munger at Crooked Timber, situating the imminent flood of GenAI-generated research publications – “greater and greater volumes of meaningless and unread text circulating for the sole purpose of individual academic careers” – within a broader context of the organisation of the production of scientific knowledge.

I know that serious scientists don’t wanna hear it, but the scientific knowledge we produce is obviously and strongly structured by our institutions. The tighter the labor market and the more artificial the metrics we use to evaluate each other (the farther from actually reading the work and subjectively evaluating its quality), the more power these institutions have. And now these for-profit corporations are setting the agenda for how LLMs will be incorporated into scientific practice.

The rules are not set by academics working as a collective, agreeing how science should be conducted under new conditions according to agreed principles, but by commercial publishers maximising their profits (and decreeing, as Munger notes, that the only ethical issue is openness about the use of GenAI), and universities setting up systems to manage and control the activities of their employees (incentivising increased output of publications, in the ‘best’ journals), and most academics going along with this because we’re all atomised individuals trying to survive within an increasingly hostile system. It’s easy to imagine these institutions conducting surveys of academic views about Open Access publication and research evaluation and the like, not in order to improve the system but to identify resistance to it and ensure greater conformity.

Anyone talking about “The Ethics of LLMs” and scientific publishing is trying to sell you something—or, in the case of Springer Nature, trying to buy something from you with your own money.

Can we extend this to ‘the ethics of GenAI and teaching’? Obviously the GenAI companies want to maximise use of their tools (hence deals with US universities and promotions aimed at students) and to make a world in which these tools are ubiquitous seem both desirable and inevitable; universities (like states) are as ever terrified of being out of step with the demands of the corporate world and attracted by the shiny new thing. And we seem to have a fair number of people aiming to make a living from inserting themselves into this circuit; maybe not wholly convinced of the benefits of GenAI themselves, but fully on board with the potential of making money from explaining to universities how they should adopt the shiny new thing and what they should tell academics to think about it. As ever when the car needs an expensive service and new tyres, I wonder about moving into consultancy work…

November 6 2024

I sit in a cafe in Innsbruck, doomscrolling, as waves of anger, fear and recrimination circulate over social media. The bright autumn sunlight illuminates the ‘Ehrenmal’ in the square in front of the university’s main building: an angry-looking bronze eagle on a triangular stone pillar decorated with the words ‘Ehre – Freiheit – Vaterland’, constructed in 1926 to commemorate the war dead and the loss of the region of Südtirol to Italy under the Treaty of Versailles (“Deutschland, Dein Reich komme!”, declared the Prorektor of the University at the inauguration).

In recent decades, the monument has been progressively modified and adapted to try to defuse its message – the dangerous conflation of honour, freedom and dedication to the Fatherland that could have come straight from Pericles’ funeral oration – and to limit the amount of vandalism, including painting the eagle pink in 2010. Most recently, the university commissioned an artistic intervention that added the word ‘welche’, in red, dripping lettering, to each face of the pillar: What honour? What freedom? What Fatherland? A fortnight ago, not long after my arrival here, many of these letters were removed overnight; presumably by ultra-nationalists who object to such questioning of traditional values. It seems unlikely to be a coincidence that a few weeks earlier the Freiheitliche Partei Österreichs had gained the largest number of votes in the Austrian general election, shadowing the rise of the far right across Europe and much of the world. Democracies voting for strong leaders who promise a return to tradition and the good old days, an end to questioning and other liberal values, and indeed to democracy. And now the United States follows them.

Exiled Thucydides knew… This really isn’t the chapter I wanted to write.

I Spy…

…with my little eye, something that looks rather like generative AI. It’s been quite a while since the last new ‘Thucydides’ quote on Ex-Twitter (where @Thucydiocy remains active, though my personal account is now dormant). Activity there has focused recently on a couple of lengthy and highly entertaining discussions of the ‘Scholars and Warriors’ quote – one person insistent that because the line was also quoted by Mustafa Kemal Atatürk then both he and William F. Butler must have been quoting the same ancient source (went silent when asked for Ataturk reference, plus fact that he was born just eight years before the publication of Butler’s book so if he did use the line it’s not exactly impossible for him to have read it in Butler), and one who appears to have an almost religious faith in the veracity of Goodreads, convinced that the Thucydiocy Bot is either genuinely a bot or is possibly a teenager. So, many thanks to an article by one Riva-Melissa Tez in The Spectator World, bemoaning the fact that even Trump may not adequately destroy democracy in the interests of libertarian Musk groupies (say what you like about The Spectator, but it does generate a lot of material for studies of stupid right-wing appropriations of the classical):

We are so much worse off than the Athenians during their similar stages of decline. Thucydides once wrote, “The Athenians, who were the most democratic of all the Greeks, were also the most prone to make mistakes, for they were always in a hurry to decide, and were swayed by the emotions of the moment.” The political satire of the poets in Athenian theaters heavily influenced the city’s political decisions, just as TikTok and the Guardian sway millions of malleable minds now. I wrote this in my last political lament in 2020, and I repeat it now — at least the Greeks had Euripides!

That quote – both the whole thing, and chunks of it large enough to be distinctive – is that rare thing, something which produces just a single result in multiple search engines, namely the article in which it appears. That suggests not just that it’s not an actual quote, but also that it’s not someone else’s paraphrase either (at first glance, I assumed it was going to be taken from V.D. Hanson’s introduction to the Landmark Thucydides or something similar, which quite regularly get circulated as actual quotes – but they are easy to Google).

The thing is, all the components of that line ring more or less true, mostly as things said by different characters in Thucydides’ account. I’m not entirely sure about ‘most democratic of all the Greeks’, but ‘most prone to make mistakes as always in a hurry’ looks like a variant of 3.42 crossed with 1.78 plus the Corinthians’ characterisation of the Athenians as always restless and active, and the point about being swayed by emotions in the moment echoes both Cleon in the Mytilene Debate and Nicias in the Sicilian Debate.

What you don’t get is all these ideas compiled into a single critique of Athens and its democracy, conveniently encapsulating a lesson for the present about why autocracy is necessary and how it’s really unfair that people stopped talking to the author at parties. It could certainly be someone’s interpretation of why Thucydides thinks Trump needs to win, but in that case it would surely appear in search results. No, either Riva-Melissa Tez has made this up herself, or – and this seems more likely to me, given that a compilation of different things that people regularly say about Thucydides is exactly what you would expect to be given – she’s asked ChatGPT for a suitable quote. A quick online search reveals that she was heavily into AI start-ups back in 2017…

Giving the author the benefit of the doubt, I have asked for a reference via Ex-Twitter, and firmly expect to be ignored. But I will update here if I do get a response. Meanwhile, I wonder whether a reflection on Thucydides Book 8 – democracy votes to abolish itself – would be at all useful as a means of responding to current events…

Waterloo

I had an interesting conversation the other week with one of my new Innsbruck colleagues about the effects of how one conceives of one’s area of study. One familiar image is that of the battlefield; certainly it’s the one I’ve come across most often, not least because Keith Hopkins used it to describe debate about the ancient economy in a much-cited piece from 1978, still being explicitly echoed in a new Handbuch der Antike Wirtschaft (von Reden & Ruffing, 2023). It’s all about a struggle against a defined enemy, strategically deploying articles and reviews to outmanoeuvre them, constantly searching for new weapons or tactics to undermine their defences; other scholars are allies or potential foes. And that of course extends to graduate students, being trained up to take their place as cannon fodder – and perhaps needing to be scrutinised for potentially traitorous thoughts. Shout-out to veterans of the Loxbridge meetings back in the early nineties, where (at least from the Cambridge side) we were expected to demonstrate that our teachers were much cleverer than your teachers…

It’s easy to see, if you take a step back, how this conception of the world and one’s place in it would then shape all aspects of professional behaviour and even self-perception; approaching every seminar or public lecture as a field of conflict, constantly looking out for ambushes, while the drive for victory always contains the possibility of ignominious defeat. It’s also entirely possible to think of different metaphors. What happens when we think of the study of the ancient economy, quite naturally, as a kind of market? Still (viciously?) competitive, but in a quite different manner, and perhaps this offers a better way of thinking about both the rise of new forms of research management (the shift from the individual artisan scholar to the Taylorised research project enterprise) and the effects of globalisation (the irresistible rise of English as the language of publication and the Anglophone university as the template for success).

I’ve suggested before that the model of collective improvisation in jazz might offer a more positive vision of what we’re all trying to do – and that’s just a specific example of a communal, collaborative, conversational approach to scholarship, in which the goal is simply to create something new by building on one another’s ideas or to play with possibilities to see where they end up or just to enjoy the exchange for its own sake. Of course – like actual jazz – this also has to take account of the real world, in which financial constraints and competition and exploitation are a constant presence (there are only a few opportunities to do this professionally, let alone to enjoy substantial rewards). But it’s surely important to hold on to the collective ideal, even if you’re one of the lucky ones who gets to be treated like a star. And, yes, I’m sure we can all think of an example of a scholarly Wynton Marsalis, trying to police the boundaries in the name of tradition.

I wonder how Richard Duncan-Jones thought about it. It’s not something I ever talked to him about, as we only interacted once or twice; he didn’t often come to events in the Classics Faculty in Cambridge, but stayed in Gonville and Caius College, teaching (I assume) Caius students, and working on his research. I think of him in this context because clearly Hopkins did, and was baffled by him: where did he stand? what side was he on? “Can you try and find out what he actually thinks about anything?” he said to me on hearing that I was going to see Duncan-Jones to talk about an issue in my own work. If Hopkins had thought through his own metaphor a bit more, he might have concluded that Duncan-Jones was genuinely neutral, manufacturing scholarly material – not weapons, but precision ball bearings or finely-ground lenses or something like that – that could be employed by anyone in their own work, without any concern about their end use. Clearly, however, he regarded every new publication from that quarter as potentially booby-trapped.

I never got anywhere with answering Hopkins’ question because I was too concerned with my own problems: I was going to talk to Duncan-Jones because a massive aporia had opened up in my research project, the question of the population of Roman Italy, and he had written the OCD entry on the topic that had neatly raised questions about all sorts of things that were supposed to have been settled. I could have stuck with the existing consensus, but I was young and foolish, and felt that every part of my dissertation needed to be equally robust, and if the population of Roman Italy could with equal plausibility be five million or fourteen million, how could we know anything about anything?

Duncan-Jones gave me lunch in college, and listened politely, and manifestly had no idea why I was getting so worked up; he had simply asked questions that needed to be asked, and if that established that we didn’t really know anything, well, that’s where we are. Manifestly he was not, as Hopkins presumably feared, slowly moving different forces into position in an apparently random, directionless manner that would suddenly be revealed as a strategic masterplan leading to a devastating coordinated strike. He studied specific historical problems as puzzles, fascinating in themselves but pretty well entirely isolated from one another. I had to work on my existential crisis in my own time.

The Duncan-Jones style of scholarship – almost pointilliste, you could say – is utterly alien to my temperament and interests; so I continued to read his publications but never made any effort to renew conversation, and have no idea of whether he read my floundering efforts in Roman demography (maybe) or my analysis of his rhetorical style as a counterpoint to Hopkins’ (probably not). Neither antagonistic nor competitive nor collaborative; scholarship as a matter of individual concentration and craft, endless meticulous tinkering and a genuine indifference to whether anyone else cares. If he didn’t have a room in college, I imagine he would have had a shed.

This resolute march to his own beat doubtless explains why it was only by accident, six months after the fact, that I learned he had died. No announcement on the Classicists email list, so far as I have seen, just a short notice on the Cambridge Classics Faculty webpage that I stumbled across when looking for someone’s email address. I imagine that a bit more fuss was made in Caius, but I don’t know.

And I’m conscious that I may be making this sound like a sad story, without any evidence that this is the case. If your goal is neither to drive your enemies before you and hear the lamentations of their graduate students, nor to be acclaimed and envied by all as the dominant figure in contemporary ancient history, nor even just to be invited to somewhere nice occasionally to give a paper, then slipping away quietly is no tragedy. There’s just the regret, one imagines, that there was always another historical problem to be explored that now perhaps will gather dust for a few decades.

Zwölf Tage im Jahr: 27ten Oktober 2024

Slept through until seven for a second night running, after a week of being lucky to get past five without thoughts of work kicking in. The gloss on this little victory is only slightly abraded by the realisation that the clocks have gone back so it’s only six; what’s important is that I have not succumbed to any after-effects from last night’s Törggelen – a traditional Südtirol autumn feast – with the Innsbruck Gesellschaft für classische Philologie: not so much the drink (two beers, a glass of the semi-fermented wine called Sturm, and a Birnenschnaps) but the massive risk of choosing Käsespätzle as my main course, which I love – and this was a very good example of the dish – but which often does not love me. No digestive issues, no gall bladder problems, only a few rather odd dreams that might in any case be put down to an evening of intense German practice, much of it involving accents which I’m still getting used to.

Dozed for another half hour, trying to find the delicate balance between nudging my brain into thinking about the chapter I have to write (by several months ago) and not getting into a state about everything. Cup of tea and biscuit in bed, catching up on news, which remains dire on multiple fronts – and having been surprised yesterday by the rolling news summary in the tram covering the ‘Tommy Robinson’ protest in London, one almost starts to wonder if there is a psychological element of normalisation through comparison – we Austrians are not the only ones with a rising tide of fascism, look at Britain/America/Germany/Slovakia/France… When confronted for example with the vandalism earlier this week of the Ehrendenkmal in Christoph-Probst-Platz at the university, a rather brilliant reclaiming/repurposing of an interwar memorial that one imagines drives right-wingers up the wall, there may be a wish to feel that it isn’t solely an Austrian crisis.

Another cup of tea to go with my muesli – a proper breakfast of cheese, ham and Vollkornbrot seems a fairly bad idea after last night. I wait for A to send me a message to say that she’s awake; she’s been having a fairly dreadful couple of days, having gone down with some sort of flu combined with the cats acting up massively, so the least I can do to alleviate guilt for abandoning her for two months (on and off – starting to get emails about my trip home in the middle of next week) is to avoid waking her up, while responding promptly as soon as she does write to make it clear that I’m not just too busy having a good time to think of her. Mostly just hoping that the shrew Hector brought in yesterday and released in the bedroom has found its way out and/or been eaten, and that he doesn’t find a replacement. There’s a pattern here; I think it was June last year in Bochum when I got a tearful phone message: “There’s a THING!!!” Which is not an easy situation to deal with on the phone.

Having had a full day of local culture yesterday, today is all about quiet: reading, listening to music, drinking tea, and occasionally trying to kick my writing brain into action, with limited success (it definitely doesn’t help that I really don’t want to write this particular chapter). Went out for a short walk around lunchtime, to take advantage of the glorious weather; along the foot/cycle path westwards up the Inn valley, until I came across a Naturschutzgebiet by the side of the river. Apparently it has at least six bat species including a rare one, so I really regret leaving the bat detector at home on the basis that by late October they’ll all be snuggled up for the winter.

My foot – the one I broke last year – has been aching a bit, either because I’ve overdone energetic tramping up mountainsides or because the weather is changing; so, took things slowly and easily, and it actually started to feel better. Passed a field of brassicas that had just been harvested and ploughed – and there were some kohlrabi and Calabrese heads in excellent condition just left on the side of the field, so that’s supper. One thing I have certainly learned on this trip is how good slices of raw kohlrabi taste when seasoned with a little salt and a glass of beer.

Composed a chord sequence in my head, with no idea whether or not it will actually work; while I can log into my jazz composition class without any difficulty while I’m here, it is proving extremely tricky to do any actual composing just on the laptop, without a piano keyboard or anything else on which to try out combinations of notes, other than laboriously inputting everything into the programme and then laboriously tweaking it again and again.

And then managed to get some words onto paper – not terribly coherent words, but more than none – before setting up the VLE so I can listen to Jazz Record Requests (possibly should not admit to evading the BBC’s attempts at limiting access to its programmes outside the UK…). Really much too much rather conventional vocal jazz for my taste – while I do enjoy a bit of traditional Alpine Volksmusik, I could really do with some less predictable rhythms and chord sequences to cleanse the palate.

Usual Sunday evening phone call to check on aged parents; it is actually remarkable how much chattier my mother has been over the last couple of weeks, as if she is rather bored of my humdrum life in Somerset but can get excited about Austria, which is interesting from a lifelong Francophile. Cooked and ate supper (reheating my attempt at a chicken curry from Friday, now adding substantially more chilli and other spices to try to offset the bland); long FaceTime chat with A, making the most of her feeling less ill and the cats being temporarily relaxed and peaceful. Early night in the hope of being clear-headed and focused for the new week; straight off to sleep – but the fact that I’m writing up these notes at just before 6 am, having already had a cup of tea and caught up on news, tells you how well that went…

Glory Days

Earlier this month, after nearly thirty years teaching in some great British universities – I mean, they’re not up to the standards of the four or five colleges in Cambridge that we all know are better than everywhere else, but they’re still pretty good – I packed my bags, held tight to my beloved copy of Thucydides, and got on a plane to Munich, from where I travelled onwards to my new station in Innsbruck. I loved my teaching in Bristol and Exeter. And that’s precisely why I seized the opportunity to grab a load more money, regardless of its source, and free myself from inconvenient bureaucratic restraints that might force me to do things I don’t feel like doing. All because students have feelings? Pah!

Once upon a time, universities were great places to be so long as you were male, white, privately educated and full of yourself, and let’s be honest they’re the only the people who really matter. There are those who suggest that now we have ChatGPT, there’s really no call for the Oxbridge supervision system’s unique ability to turn out glib, superficial bullshitters. I ask you! Can Generative AI drink seventeen pints of wine, smash up a restaurant, assault a couple of Spanish EFL students and then next morning still impress a socially awkward college tutor with its boundless self-confidence?

Where did it all go wrong? There are many whose nostalgia is merely a cri de coeur for their lost youth, vigour and hair, who therefore idealise and idolise an imagined version of their own university days. They mistake their own feelings of personal failure for a cultural diagnosis. No, universities’ historic strengths came from giving arrogant aristocrats the opportunity to drink, whore and fight occasional duels to their hearts’ content while a few weedy types coughed syphilitically over dusty books, and when academics who succumbed to bourgeois morality by getting married were sent off to spend their lives preaching every other Sunday to five farmers, their families and a couple of curious sheep in darkest Herefordshire. When we lost that, it was over. Not to mention all that bloody science stuff. If we’re going to spin five hundred words out of pandering to the anti-intellectual prejudices of elderly grumps and smarmy would-be SPADs, let’s get serious about it.

Today’s students? Are they reading Plato’s Republic in the original with Hegel and Schelling, then falling passionately in love with the mother of the children they’re supposed to be tutoring, getting fired, walking all the way from Bordeaux to Stuttgart for no terribly good reason, undergoing abusive psychiatric experiments and spending the last thirty years of their lives in the spare room of a local carpenter, all the while writing classically-influenced poetry?* The fact that they clearly are not tells you all you need to know about modern British university education. Extra time for exams? Refusal to undergo public shaming for disappointing degree results? Widening participation? COURSEWORK?!? No wonder they’re all fucking liberal snowflakes. Excuse me while I faint onto my chaise longue.

I mean, if Cambridge can’t resist being dragged into the later nineteenth century, who can? Now, please excuse me while I monetise my atavism.

*Why yes, I did hear a very entertaining discussion of Hölderlin the other day…

The Sound of Silence

Yesterday afternoon I invented a new index for evaluating blogs – so new that I haven’t come up with a decent name for it yet – that is the ratio between posts and authors over a given time period, most likely a year. So, a single-author blog with one post a month has an index of 12; one with one post per week has an index of 52 or so. This potentially allows comparison between single-author blogs and collective ones like Crooked Timber or Lawyers, Guns & Money – huge rate of posting, especially with the latter, but that’s driven by there being five or six really obsessive posters and assorted others so the index is probably just a couple of hundred for LGM, and under 50 for CT these days.

The really neat thing about this index is that it weeds out those ‘blogs’ that are really just a sort of web magazine, with an editorial group doing the reading and commissioning, and scores of authors who typically publish just a single piece on the platform. The nearer the index approaches 1 (e.g. 50 articles per year from 48 different authors), the less you can claim to be a serious blog, either because you’re not posting very much or because you’re not really a blog at all but just trying to steal the kudos and rebel authenticity of those of us who keep ploughing our lonely and unproductive furrows…

All of this is just a prelude, or rather a reminder that I need to keep this place going even while I’m thoroughly absorbed with settling into a two-month stay in Innsbruck, practising my German, doing a bit of teaching and (in theory) lots and lots of writing, and (so far) drinking too much coffee. The Institut für Klassische Philologie und Neolateinische Studien, who have found me a corner of an office with a magnificent view of assorted churches and Alps, have a common room with coffee machine and someone who delights in baking biscuits and cakes for everyone, which may prove unfortunate for my waistline.

On Wednesday evening, to get finally to the point, I dropped into a paper in Reading by my former Bristol colleague Genevieve Liveley: ‘The Silence of the LLMs’, a title that demanded a paper be written around it if ever I saw one. Genevieve has long been working at the cutting edge that is, crudely, gently introducing computer scientists to humanities research that is directly relevant to the problem they think they’ve just invented. ‘Silence’ is a really nice example of this. Its multiple functions – hesitation, deferral, evasion, disapproval and so forth – are a familiar theme in literary studies, but it has apparently only just occurred to people in GenAI research that silence might be meaningful rather than just an absence.

You can get ChatGPT to fill in the gaps of a Sappho poem (not terribly well) or propose a reconstruction of a fragmentary inscription, but you have to programme a speech generator to pause ‘realistically’, and doing this too much tends to break it. More silences – putting the onus on the reader or listener to fill in or interpret the pause – might make the machine seem more human (even I, committed ‘AI’ grouch, fall into the habit of saying “it’s thinking” when a programme stalls for a bit), but not if you have to tell it to. Perhaps an entity made up almost entirely of words experiences silence as a kind of momento mori…

My thoughts turned, predictably, to what could be termed GenAI’s performance as a historian: its outputs consistently exclude the possibility of gaps, silences or ambiguities in knowledge of the past, always presenting a seamless and confident account. This is a key reason why, as my students have commented, “it doesn’t write like a student”: no qualifications, no hedging of bets, no acknowledgement of uncertainty or of different, equally plausible interpretations.

I think I’ve mentioned here before (it is a key part of my personal myth, after all, such that I trotted it out for my welcome piece on the U. Innsbruck webpage, so I must have done…) that a student once bought me a mug decorated with “The simple answer is… we just don’t know”. Sometimes, the correct answer in history is a blank. But that blank is not merely empty; when we move on to the more complicated answer, we start to delineate the shape and character of our ignorance, to explain the reasons for its existence, to canvas different ways of estimating its contents, to establish its potential significance.

This is not the ChatGPT way. The simple explanation is that most of its training data in historical topics consists of crude school textbooks and popular histories that likewise erase all trace of uncertainty in favour of confident description and assertion, and resolutely refuse to get involved in arguments about interpretation of fragmentary evidence. But surely its data also include proper academic research, where uncertainty is highlighted – in which case, why is it incapable of imitating that, even when explicitly instructed to mimic a proper academic historian in its output? Trying to compel it to support its claims with evidence, let alone to contend with contradictory evidence or rival interpretations, is another way of ‘breaking’ it, at least in the sense that its inadequacies become more and more obvious. Of course the LLM is incapable of historical analysis – but it can’t even do a passable imitation of historical analysis.

That’s not to say that the output never admits to ignorance; but those admissions are a programmed response to situations where it’s accused of self-contradiction or told that its answer is wrong, rather than anything one might call a genuine admission, and its next step is either to offer a new, equally confident and unequivocal answer or simply to say that it doesn’t have access to the relevant data. It’s a binary world, in which either there is data and hence clear answers or there is insufficient data and hence no answers; there’s no sense of the possibility of ambiguous data where competing interpretations need to be evaluated in different ways, or that insufficient data might still convey meaning, or that the gaps themselves may be significant – which is of course the world of the historian.

Which brings me back to Genevieve’s talk; just as the problem of LLMs and silence reflects in part the fact that computer scientists hadn’t hitherto thought much about silence and its complexities, so perhaps the problem of LLMs and historical understanding reflects not only the training data but also the programmers’ worldview, that if there is data then there will be a clear answer to any given question that GenAI will produce, and the only alternative is an absence of data. The tool is designed to cut through complexity and sweep away ambiguity, even when complexity and ambiguity are the point; a good answer in its terms is always a definite, unequivocal answer, and hesitation or qualification make for a bad answer. This isn’t just a matter of the nature of the database, but of how it has been trained to use it.

Whereof one cannot speak, thereof ChatGPT will say something anyway.

Qui Sème le Vent Récolte le Tempo

One is never strong enough to be sure of always being the strongest.

Bernard-Henri Lévy, Jean-Jacques Rousseau, or Thucydides? It’s definitely BHL as he’s used the line in various publications in both French and English, as well as on Ex-Twitter last week (link); sometimes attributed to Pericles and sometimes to Pericles as reported/paraphrased by Thucydides (so, no immediate need to search Plutarch too, as otherwise might be required for Pericles quotes).

Anything involving strength, power, implied conflict etc. instantly looks Thucydidean, because of the Melian Dialogue and the tradition of reading him as a prophet of Realism in international relations. But BHL’s wording doesn’t show up in any online searches (besides a few other people quoting that tweet), so he’s not citing any published English translation. But of course he’s French, and the first time he used this quote (an article in Le Point back in 2007, so far as I can ascertain, reproduced on his webpage: link) it was in French: “…la certitude que, comme disait Thucydide à propos de l’Athènes de Périclès, personne n’est jamais n’est assez fort pour être sûr d’être toujours le plus fort…” But that version doesn’t show up anywhere else either. Caveat: I’m not sure if all French translations of Thucydides are online, so can’t be 100% certain that he’s not quoting one, but on the face of it this seems like his own version – whether translation, paraphrase or completely made up but attributed to Thuc as the sort of thing he might have said.

Now it gets tricky: is there a line in Thucydides that looks as if it could be translated or paraphrased like this? How can one prove a negative? As I’ve discussed on here before, there are still people who insist that Thucydides DID write the ‘Scholars and warriors’ line because I can’t demonstrate that he didn’t. And there is of course an awful lot of Thucydides to check… BHL’s mention of Pericles potentially narrows things down, as (spoiler alert!) Pericles dies halfway through Book 2. I don’t see anything in any of Pericles’ speeches that looks much like this idea, though he did have things to say about power.

What the line does strongly remind me of, however, is the Melian Dialogue; not just because it’s all about power and inter-state relations, but because the Melian representatives do say to the Athenians words to the effect of: you won’t always be supreme, and then just wait for the blowback. (To be precise, it’s 5.90: “There is advantage in your preserving the principle of the common good… This principle is proportionately in your interest much more than ours, given the massive retaliation you would face as an example to others should you fall from power”, trans. Hammond).

So it’s possible that BHL is vaguely recalling that passage, and for some reason dragging Pericles into it. But there is another possibility: that he’s actually misremembering a bit of Rousseau instead. Du Contrat Social, opening words of Book 1 Chapter 3: “Le plus fort n’est jamais assez fort pour être toujours le maître.” Pretty close to BHL’s “personne n’est jamais n’est assez fort pour être sûr d’être toujours le plus fort.” Certainly the same idea, albeit R’s version is rather better written. Of course it’s possible that Rousseau was channelling the Melian Dialogue here; he doesn’t make any reference to specific sources for his claim, but it’s clear from other writings that he knew Thucydides and sometimes drew on his ideas. I am waiting to hear further thoughts from French colleagues…

So, according to the rating scale of Thucydides quotes that I’ve suddenly realised I should long since have invented, this looks like a 3-Kleon: plausible paraphrase of genuine bit of the text, but with misleading elements and of course no proper reference. (‘Scholars and Warriors’ is the gold standard for a 5-Kleon quote, not just completely wrong but also completely anachronistic).

Why does any of this matter, about from my obsessive pedantry? Well, a major aspect of Thucydides’ contemporary significance is his image as a particular kind of author – analytical, illusionless, focused on issues of power. This is partly based on various quotations, including misattributed ones, and partly works in reverse, as this image then works to legitimise a load of dodgy quotations as being the sort of thing that Thucydides might or should have said. He’s a convenient author to attach to some quotes as a more authoritative source than the actual author.

[This is basically just a reproduction of a thread I posted on Bluesky for Thucydides Day in #ClassicsTober24, but I wanted to keep a record in case I need to check this quote in future.]

Best of All Possible Worlds

I’ve just got back from a round trip to visit my parents – seven hours of driving in intermittent downpour in order to have lunch and complete assorted tasks involving stepladders, heavy lifting, wading in ponds to move water lilies and so forth. And then coming away with two bottles of sparking wine, a jam jar full of ramshorn pond snails, a complete set of Tintin in French (they’re having a very, very slow clear-out, and for some reason my philistine brothers weren’t interested in these) – and one of the funniest ‘sliding doors’ possibilities for my life I’ve ever imagined. I’m going to tell the story rather differently from the version my father told me this afternoon, as his lacked narrative coherence…

The one tangible success of my entire musical career came back in the early 1990s, when a selection of my self-recorded songs – ah, the wonders of the old Tascam four-track recorder – won a competition in the student newspaper. I must have shared the cutting with my parents, or at any rate my father, who was delighted by the fact it made reference to the homemade equipment in my recording setup (he’d built the Great British Spring Reverb for the sheer engineering joy of it, as well as my guitar amplifier), and tickled by the lyrics that were quoted. He therefore shared these with his great mate – an Oxford-educated, crossword-solving, OED-contributing polymath called Mike Barnes, who must have endured a lot of updates on my intellectual exploits over the years.

Mike was sufficiently interested and amused then to mention this to an old university acquaintance of his, a Rhodes scholar from Merton College with whom he had boxed once a week, who had eventually gone on to a successful musical career. Yes, this is where it suddenly becomes clear why my father had suddenly remembered the story; his best friend had been in regular correspondence with the recently deceased Kris Kristofferson. Even more hilariously, Kristofferson had actually liked my lyrics, and wrote back to Mike that the lines in question really spoke to him, and if his (Mike’s) friend’s son would like to send over a tape of songs he’d be happy to offer some advice.

My father never passed on this message to me – he’d obviously decided that I would be much too focused on exploring the economy of Ancient Rome to consider such a thing. And the very little I knew of country music at the time was mostly unfavourable, and the songs I was writing were striving in a completely different direction, and I imagine Kristofferson would have been entirely unimpressed. But on the other hand I can appreciate the musical craft in this genre, and the emphasis on lyrics and story-telling, and the fact that someone like Kristofferson could happily build a song around the line from Doctor Pangloss about the best of all possible worlds – and so there is a conceivable timeline in which Neville then abandoned the University Library for Nashville, not to become a star – that cut-glass English accent was never going to cut it – but to churn out a respectable number of hit songs, ideally in a less cynical vein than Sven in Questionable Content.

Obviously Kristofferson comes out of the story best; not just the Rhodes scholar and boxing Blue part of his legend, but the fact that he sent regular birthday and Christmas cards and letters for decades to the bloke with whom he’d sparred for a year, regardless of his stardom, and was perfectly willing to offer genuine advice to the son of a friend of this friend. And I get to say that a song-writing great once liked one of my verses. But what my father thought he was playing at…

Most Likely You’ll Go Your Way (And I’ll Go Mine)

It remains a source of anxiousness and confusion that I’m not having to worry about teaching this year. I dreamed about it last night, which suggests that I’m worrying about it anyway… I’m not sure if it this relieved or accentuated by continuing to think about the whole GenAI thing, not least because I was making a presentation to my department today about my project and possible responses (both to the technology, and to the university’s chosen approach). Whether it helps or not, I did find myself composing what I would be writing in my module handbooks for this year, were I writing any module handbooks…

Assessment

Following the university’s new guidance on the use of AI and GenAI, the core assessment for this module is ‘AI-Supported’. What this means is that, if you make use of any sort of AI tool in developing your coursework, you will not be penalised for this provided that you acknowledge all such uses, using the new cover sheet, properly reference any text that has been generated, and confirm that you are aware of how to use AI ethically and responsibly.

If I had a free choice, however, I would describe this assessment as ‘AI-Tolerated’. That is to say: you will not be penalised for using any sort of AI tool if you properly acknowledge this, but I would recommend great caution; there is a substantial risk that you will be penalising yourself, as coursework based heavily on GenAI output tends to be at best passable according to our standard marking criteria. Making effective use of GenAI requires considerable knowledge and critical understanding, to recognise that most of its outputs are problematic, dated, tendentious and sometimes fictional. Using AI summaries of articles or chapters, even if you can trust the summary, means that you are not in fact learning any of the skills of critical reading and interpretation. Finally, it is open to question whether there are ethical and responsible uses of GenAI; I myself would not readily sign a statement to the effect that I am aware of how to use AI ethically and responsibly.

Developing a critical understanding of these tools is essential, and so if you are really interested I suggest you try the ‘AI Integrated’ option for coursework. Otherwise, be very cautious and sceptical. We will be exploring some of these issues in class…