Hacker News Reader

Comment by openrisk | Hacker News Reader

> Archeology is not code nor is it physics.

Indeed, but after scanning this article that pulls in all those pieces of indirect evidence I wondered whether some type of structured knowledge database (that encodes the innumerable pieces of historical information that are known, tags them with confidence levels etc.) would not be useful to advance research in such domains.

Something like a large collection of RDF triplets against which you could run a query like "Given this new data point how (more)likely that Alexander the Great's tunic is identified in a royal tomb at Vergina?"

View on HN

To me it sounds like it could (and likely would) backfire, by replacing judgment with numbers. Who is giving the confidence score? What confidence score does each confidence score receive? Why are those scores more valid than the expert in that very narrow domain? If that expert is the one giving the scores, are they not just gatekeeping? Et cetera. I don't want to see researchers rewriting their papers because their cumulative source score is 68.17, and it should be 72.5 or higher.

also, there have been points in time where established archeology was wrong, and this seems like it would produce a bias towards what we currently think is true.

for example, theories on how the Polynesian migration came to be are still in flux, to the point where one theory was attempted to be proven by actually sailing to the different islands using only traditional wayfinding.

> replacing judgment with numbers

I would phrase it otherwise: supporting judgement with numbers. Its not about altering conclusions, but making more transparent the factual basis and associated reasoning from which they are derived.

The analogy would be trying some exotic food and having a list of ingredients. Yes, good to listen to a local as to how it tastes (and whether it cures all diseases), but if the indication is: 50% sugar, thats a data point worth knowing.

You would attach names and dates to the numbers, as with any scientific publishing.

I think that, effectively, the corpus of research papers and citation links is this knowledge database. It isn't structured the way I would structure it in postgres but it seems to be working quite well for the professionals in this field.

I know there have been some interesting finds when an archeologist has dug up a site report from the 1840s that had long laid ignored by academia but these are quite rare occurrences and the scale of people involved here (when we're talking about something hyper specific) is so small that they can probably just sort it out by talking to one another.

For the outside public such a neatly tagged database might be helpful if someone outside of the circle wants to independently research a subject in depth but, honestly, these folks are pretty open to questions and discussions so if you're extremely interested in Gobekli Tepe or some such there's someone out there who is happy to start a conversation with you.

> the corpus of research papers and citation links is this knowledge database

yes, I think so too. In the typical fashion of "pre-digital" information management systems it is extremely economical in the way it encodes things, with statements like "X is true as shown \cite{Y}" etc. But...

> but it seems to be working quite well for the professionals in this field

what prompted my comment is exactly the fact that didn't seem to work that well in this case :-) (nb: I am not remotely an archeology boffin, just triggered by the adversarial language of the paper).

In more quantitative fields people talk about reproducible research, here its more a question of whether similar fields would benefit from "reproducible chains of reasoning".

> it seems to be working quite well for the professionals in this field

That is the universal response to new technology: What we're doing is working fine! What they are saying is, 'everything we've accomplished has been with the old technology'.

I promise that was heard from engineers and architects encountering CAD, from cavalry asked to give up their horses (the conservative urge is so great, many died charging machine guns!), by literary scholars presented with computerized tools, .... it's always the same. One person who installed the first email systems for many businesses told me that, over and over, people would say 'our paper memos work fine - this is just technology for technology sake'. They meant, 'everything we've accomplished, we've done it with paper memos'.

New technology lets you do old things much faster and/or lets you do new things you couldn't do before - new things you didn't dream of doing, and as people discover uses for it, new things you won't know about for years.

And the universal argument that people pushing tech are making boils down to 'I don't understand your field, or the particular needs of it, but I'd like to sell you a process that I invented. I'm not going to be held responsible for any bad consequences of you adopting it.'

Unsurprisingly, people tend to resist this sort of thing.

Sometimes the local maximum people are stuck in sucks, and they need a shakeup.

That shakeup will not be well received when it comes from a complete stranger, who has no rapport with the community, with zero skin in the game.

I agree 100%. The number one issue is buy-in, by the leadership and by the users. Without it, don't waste your time.

Buy-in requires their input and demonstrable benefits to them.

Something like this, perhaps? https://digitalculture.uchicago.edu/platforms/ochre-overview...

It ain’t what you don’t know that gets you into trouble. It’s what you know for sure that just ain’t so.

That would introduce you to a world quite similar to psychology's obsession with p-hacking to prove confidence where there should be none.

Reply on news.ycombinator.com