Datalore
BLOG FIGHT!
Shoulders Of Giants
This is the post in which I’m finally going to have to fully acknowledge my debt to Alina Marsfelder and her Vaka Rangi blog. I read the whole thing though just after she (essentially) finished it a few years back, and there is precisely zero chance that the insights and approach contained within haven’t influenced how I view and talk about Trek – indeed I cited VR just on Thursday.
My respect for Marsfelder’s work doesn’t mean I’m trying to ape her – at least I very much hope that’s not what I’m doing. Once I finish a first draft of an IDFC post, though, I make a point of reading through her own essay on the respective episode (when one exists – I’ve got Voyager and Enterprise all to myself), just to check I haven’t unconsciously echoed something she originally put into my head. Sometimes, as you may have noticed, this leads to me linking to her blog, when I realise I’m either running parallel to her position, or responding to a point I first saw her make.
With “Datalore”, though, we reach something different – the first episode I’m covering for which I remember Marsfelder’s comments so clearly, and think so highly of them, that I don’t see any point in even trying to avoid using them as a starting point. Especially as I’m so interested in the choices writers make when choosing names. So: the only way to kick this analysis off is with Marsfelder’s observation that by calling Data’s brother Lore, the gathering and processing of information is being placed in direct opposition to the tales we tell to and about each other. Further, by making Lore a gleeful accessory to mass-murder, the show immediately picks its dog in the fight it has just arranged. According to the episode, it’s not just that the rational, dispassionate accumulation of information will defeat and exile the unreliable histrionics of anything so mercurial as a story. It’s that it should.
How utterly brilliant and damning an observation is that? Also, though, how much more perfect a set-up could there be for conflict between Marsfelder, with her background in cultural anthropology, and myself, a statistics lecturer so unable to leave his day job behind I’ve already written two utterly ludicrous pieces on the use of probability in Trek? A slapdown between data and lore? ACTIVATE PUNCHING!
Except, obviously, that’s not what’s going to happen. I’m going to end up diverging quite a bit from Marsfelder’s position, but I’m certainly not going to do it by dumping on the practice of passing of information and tradition through fiction. Doubtless there are mathematicians who would argue oral tradition is a poor substitute for the chi-square test – I can say with sad certainty that there is essentially no view so self-evidently stupid or vile that you can’t find a mathematician who’s signed up to it – but I refuse to be one of them. I don’t just dig Trek because it exists on film.
What I don’t agree on, though, is that this ridiculous, damaging war between the story and the spreadsheet can count “Datalore” among its battlefields.
To be clear, I have absolutely no problem believing Roddenberry would have been happy with a data vs lore, even if – as Marsfelder notes – the name preceded the decision to have Lore be Data’s evil brother (the latter done at Spiner’s behest). Roddenberry’s love of entirely logical characters, and the way he positions them as superior to those acting emotionally, is sufficient proof of that (though this is complicated by the fact, y’know, the dude wrote fiction). The point where I separate from Marsfelder isn’t the idea that data and lore are being put in opposition here. It’s the idea they’re being set up as opposites. This is not a story in which the two methods of information propagation are treated as a dichotomy. Instead, it’s a story in which data is specifically defined as being lesser than lore, in the sense that data is what you get when you take lore, and tear something away from it.
Brothers On His Brother
There are two indications that this is the case, both (unsurprisingly) involving Data himself. The more immediate of the two is the fact that Data exists only because Lore proved “too” human, and had to have certain aspects deleted before he could be considered acceptable. The second comes before Lore is even found, when we learn Data contains all the memories of the Omicron Theta colonists. The script is more or less explicit that Data would rather those memories had included their owners’ fundamental human experiences; that something was discarded that he would rather had been retained. The process of reducing their life stories to a simple set of recollections has evoked a sense of loss, even in a being as ostensibly unemotional as Data.
The implication here isn’t difficult to tease out: to replace lore with data – equivalently, to swap out a story for a simple factual summary – is to consciously downgrade the degree of humanity carried by the information stream. The process of moving from the full human experience to a simple list of what humans can do and have done carries with it an incalculable and unavoidable loss. Even the planet itself backs this position up – it’s obvious that Data’s knowledge of the colonists’ final days cannot possibly make up for what was lost with their deaths.
This isn’t to argue that data doesn’t have any value; that would be just as ridiculous as arguing there’s nothing useful carried within a folk-tale. It does mean though that me must be clear-eyed about what precisely we do and do not retain when we map people onto a data frame. Watching “Datalore” reminded me of a brief exchange I had with David Brothers over at the now very sadly defunct 4thletter!. The issue at hand was the degree to which statistics could possibly capture the complicated reality of people’s heritage. Brother’s argument (as I remember it, I’ve been unable to find the original exchange) was that even a cursory glance of the story just of himself and his own brother revealed a complexity and richness that simply can’t be broken down into, say, a tick-box ethnicity question on a government form.
Clearly, this is an entirely fair point. But it’s also one that any even halfway decent statistician is fully aware of. It’s absolutely true that, in almost every circumstance, acquiring data requires a loss of pertinent information. How else can we collect and store what we most require in a sensible amount of time and space? We even have a relevant concept here: the “sufficient statistic” – a summary of the collected information which doesn’t actually toss out anything we’ll need later on. Much of the time, the summaries we find ourselves needing to apply aren’t sufficient. Even something as simple as calculating an average to summarise your data requires discarding information, because usually you can’t use an average to go back and find the original values it was calculated from. We know the average of two and ten is six, say, but if I tell you the average of two numbers was seven, you can’t possibly work backwards to the specific two numbers I started with.
So yes. Statistics involves – has to involve, was always intended to involve – simplifying or even excluding information. It’s not the near-inevitability of such losses that’s the problem. It’s the idea that such losses don’t actually matter. That there exists no difference between what we didn’t (or even couldn’t) keep, and what was worth keeping. This makes Brothers example particularly pertinent, of course. The ethnicity of a person is a rather more complicated and politically loaded and historically contingent issue than, say, whether they’re right-hand dominant. Over-simplified ethnicity data can certainly do damage when in the wrong hands. Sometimes information is discarded not because we lack the time or space to find and hold it, but because the information makes people uncomfortable. It complicates their worldview. It forces them to see people, not numbers. So when your goal is checking, say, whether universities are doing enough to attract and maintain a diverse student body, grouping together distinct ethnicities may be a necessary evil [1]. When your goal is to make a political point about how many non-white people now live in UK to justify draconian immigration measures? That’s something else.
There are plenty of other examples of how this becomes a problem. Plenty of people who claim the stories of those different to them is “unpersuasive” or even “manipulative”. People who insist they just want numbers, but are saying that because they know numbers can’t argue back when you decide to dismiss them. These people prefer data to lore not because stories can’t tell us enough about the world, but precisely because they can tell us so much.
Beyond The Valley
How does this relate to “Datalore”? Let’s return to the inhabitants of Omicron Theta, and their own rejection of Lore.
First of all, let’s talk about what that rejection doesn’t represent, which is some kind of uncanny valley response to Lore’s proximity to humanity. There are a couple of reasons to think that’s the wrong take here. For one thing, the idea of the uncanny valley is deeply caught up in the visual experience – how close to human a replica of humanity actually looks. Lore’s facial tic aside (something the episode demonstrates is easily added or subtracted), the whole point here is that the two androids are visually indistinguishable. Furthermore, Masahiro Mori’s original concept included the Japanese kanji symbol for valley, 谷, for a very good reason. Mori’s thesis was that as a robot became more human-like, actual people would first feel more and more empathy toward it, then become revolted when it reached a stage of being close to seeming “real”, before reversing course again when the robot became essentially indistinguishable from a human being. The discomfort people feel with close approximations to humanity exists inside the valley between two peaks of sympathetic reaction.
Put simply, then, there’s no reason Data would represent an escape from the valley, and in fact we could argue he’s more likely to be found there.
This isn’t about the colonists being discomforted by a robot appearing to be too close to humanity. It’s about the fear that Lore, through his fundamentally different experience of life, might (and probably did) begin to demand those experiences be taken into account. He could tell stories about himself that those around him weren’t ready to hear. So they did what so many people do, when they don’t want to face the reality of lives led differently, and how those differences might necessitate changes in their own actions. They demanded that what they were told should be “more scientific”.
They demanded Data.
Ghost Stories
The results of these demands ended up being profoundly ironic. It’s not just the brute fact of the colonists being killed by a creature so bizarre and so evoking of wonder (as demonstrated by the heart-breaking children’s pictures found in Soong’s lab) that it might as well be a creature torn from ancient myth. There’s the fact their betrayal of Lore – either demanding a sentient being be dismembered or rendered comatose, depending on how you map our terms onto android life – resulted in them generating more lore, via the creation of a “good” twin to Lore’s “evil” (to adopt the problematically over-simplified framing this trope requires). The Omicron Thetans end up responsible for spawning a doppelganger. Yes, it’s inverted in the sense the malevolent twin predates its opposite, but still; Lore is so lore Aaron Mahnke covered the concept in an episode of, naturally, Lore.
The colonists not only insist on not having to listen to Lore’s story, they insist on being able to overwrite it. In doing so, they end up generating a ghost story with themselves as the haunted victims. They recast Lore as villainous simply by dint of his existence opposite Data, and Lore plays the part to the hilt. It’s worth noting that there’s never any explanation given as to exactly why Lore summoned the Crystalline Entity to Omicron Theta. He cites “revenge”, presumably because he clearly gained nothing else from getting of hundreds of people eaten by a creature that couldn’t rescue him from his underground prison. As weak as that it is, though, it still basically makes sense – indeed a celestial being massacring everyone for their sins is precisely what you’d expect a being named Lore to consider appropriate vengeance.
Rather less explicable is his plan to feed the entity with the Enterprise crew. The Crystalline Entity is both a great concept and a great visual, but it’s impossible to understand what precisely its “gratitude” is actually worth to Lore. “I owe it for eating some people, so I’m paying it back by helping it eat some people”. Lore is evil simply and entirely because that’s the role assigned to him. The role the colonists wrote for him.
We can take this line of reasoning still further. Through their dismissal of Lore, the inhabitants of Omicron Theta ultimately become in effect lore themselves. Not just in the sense of being the protagonists of a self-penned ghost story, nor in becoming the heart of a mystery story fated to be passed around the Federation. Their small number and unthinking reaction to Lore’s creation pushes them into the of pitchfork-wielding villagers, demanding the scientist’s human-yet-not creation be destroyed. I guess when the Crystalline Entity scoured the planet to bedrock, it got rid of the burnt-out windmill that must have stood somewhere near Soong's lab.
Just as they become lore, however, they become data too, as mentioned above. Their memories are shorn of experience, and transferred into a lonely, uncertain android. Their demand Lore be reduced to Data has reduced their own stories, their own lore, into mere data for Data to store.
Which effectively brings us full circle. To summarise, then, this episode is absolutely not about the superiority of data to lore. It’s entirely the reverse. Data is portrayed here as what you get when you can’t handle lore, when you find it necessary to cut elements out of a story – or ignore it entirely – in order for it to no longer have the capacity to affect you. Lore himself might be unashamedly evil here, but the original sin here is clearly that of the colonists, who demanded they not have to listen to anyone else’s story, and in the process rewrote their own into a tragedy.
So, given all the fun and chewy material I’ve been able to spin out of this episode, do I also disagree with Marsfelder’s conclusion that “Datalore” is awful? Actually, no. In terms of the quality of the episode itself, I think she nails it completely. We may totally disagree on the nature and quality of the subtext, but the text itself is clearly awful, from Lore’s ridiculous twitch to the baffling decision to pour out stupidity pints (with a sneerful chaser) to the entire adult crew so once again Wesley can save the day. The absolute best I can say about the actual experience of watching this episode is that it’s kind of fun watching Spiner chew the gel lights.
And yet, as we’ve seen, this genuinely could have worked, literally ancient tropes and all. The information was very much there. Roddenberry simply chose to throw away the wrong things.
Then again, Roddenberry never claimed to be a statistician. What he’s creating here is lore. And despite his obvious limitations in this area – limitations very much on display in “Datalore” – it’s clear that he succeeded. He may have let Brent Spiner cast Lore as an enemy, but it’s clear that lore itself had a very different meaning for him.
It was his legacy.
Ordering
2. Datalore
[1] I describe this as necessary because, when you're trying to prove your student population doesn't contain nearly enough non-white students, presenting numbers for multiple non-white ethnicities actually dilutes the message of white dominance. This is both because people find it easier to compare binary counts than counts across many groups, and because smaller group counts come attached to greater degrees of uncertainty, making it easier for people desperate to ignore you to pretend they're ignoring you for statistically sound reasons.