The Wikipedia Knows Nothing

Can someone who looks up a topic on the Wikipedia be said to have access to knowledge on that subject?

Questions about what constitutes knowledge are the purview of a field of philosophy known as ‘epistemology’, and the most general formula for understanding what constitutes knowledge is justified true belief. Deriving from Plato’s Theaetetus dialogue, the core of the idea is that we have knowledge when we hold a belief about the world that is true, and when we also possess a legitimate justification for that belief. I have reasons to question this model, but this will have to wait for another time. For now, it is sufficient to ask whether anyone relying on the Wikipedia can be said to possess justified true belief.

It is clear that there a good reasons for doubting that someone whose only source was the Wikipedia can form a justified true belief. In the first place, any time we refer to the Wikipedia, we have no way of knowing whether the information recorded there has been mischievously altered by someone, effectively at random. The open editorial policy means that we can never eliminate this possibility, which gives us reason to doubt the accuracy of anything read there. Furthermore, while the content policy means that any claim posted must be referenced to some other source, there are limited standards of quality control regarding this referencing: it is perfectly possible to accurately reference inaccurate claims, for instance. So we can also doubt that Wikipedia, by itself, provides any basis for legitimate justification. Indeed, the only times we can actually be confident that the Wikipedia has valid content in an article is when we know that topic from some other source. At which point, the Wikipedia has ceased to be a provider of knowledge at all, and has become merely an aide-mémoire; a means of reminding.

How does the Wikipedia differ from, say, a traditional encyclopaedia? The construction of a reference work of this kind involves an editorial team selecting experts in various fields to write articles that summarise topics relating to those fields. Clearly, the knowledge so recorded can still be wrong – examining early twentieth century encyclopaedias reveals the extent at which our knowledge has been substantially revised just in the last century – but where such a book is correct, we have at least a strong claim to justification: the expert who drafted the article text was selected precisely because of their knowledge of the field in question. We can, in principle at least, possess justified true belief if we start from a conventional encyclopaedia to a degree that cannot apply to the Wikipedia. The decisive point isn't the truth of either source: it is the lack of reliable claims to justification on behalf of the Wikipedia.

What this makes clear is that the way an encyclopaedia is supposed to record knowledge is based upon an appeal to the authority of experts. The problem with the Wikipedia is that the editors are in no way guaranteed to have expert knowledge, and neither are they necessarily qualified to distinguish expertise from its alternatives. This does not make Wikipedia useless: it functions as a giant ‘dial-a-geek’, whereby you can get an informed opinion via a random unknown nerd or cluster of nerds. It is likely to be correct when the information is straightforward (historical dates, for instance). But beyond trivia, its epistemological claims become increasingly sketchy. It certainly could contain factual information – but you can never be sure that what you are looking at does, if you don't have some other point of reference.

The Wikipedia knows nothing, or rather, someone using the Wikipedia cannot know anything from that alone. Whoever refers to a topic on the Wikipedia cannot be said to possess justified true beliefs (i.e. knowledge, conventionally construed), because no-one who edits the Wikipedia has been credibly selected for expertise, creating a gap in justification. Wikipedia editors have self-selected based on personal interest – which is why there are so many articles about (say) Star Trek and videogames. This process could still convey expertise – it frequently does about as well as any other encyclopaedia! – but we are never certain that it has done so in a subject we ourselves know nothing about. Even when Wikipedia provides correct information we cannot know it has done so, and thus can possess no justification for claiming that we do. I say this as someone who uses and even edits the Wikipedia on a regular basis, and who appreciates its remarkable virtue as a public database. It is not that particular resource that I am questioning, but the very idea of databases as a means of knowing. Kn0wledge, whatever it might be, becomes tainted with doubts when it is compiled using the kinds of techniques that lie behind the Wikipedia – and this in turn raises interesting questions about all our knowledge.

This enquiry continues in Factual Knowledge, next week.

Comments

[ptermx]

Isn't the same true of any source of information? When I read a peer-reviewed paper in an academic journal, I still have to trust that the authors, reviewers and editors were all honest and competent. Yet we know academic peer review is not all it's cracked up to be. Reviewers can be biased, or review papers they don't fully understand or sometimes don't even read them fully. Sure, it's better than Wikipedia, but still imperfect. The only real assurance you've got is to go check it all for yourself. But if you have to do that, what use was the information source to begin with?
Well ... actually, it could be that the only reason you bothered was because you read the source and doubted it. Even if the information was wrong, the source had value in prompting you to make your own investigations.
The commonly held view seems to that the ethical standard to which information sources should be held is one of whether they are accurate (true) or not. I wonder if we should instead value sources by the extent to which they prompt the imaginations of readers to conduct their own first-hand investigations.

[Chris]

Many thanks for your comment, ptermx!

Firstly, you're absolutely right - this question of trust is crucial when we think of knowledge in terms of propositional claims. This, in fact, is the stepping point for the second of these three short pieces (that form one enquiry) - so I think when you see the other two parts, you'll appreciate just how much I agree with you here! :)

But my ultimate position is that we might have reasons to reject propositional knowledge as knowledge anyway - more on this next week.

It's a fascinating idea to recognize value in a source according to whether it prompts further investigations! My instinctive response to this is that this is only one of a suite of values we can have in regard to sources, and that some of the things that possess this value would be rather poor in some other values! (An obviously badly-constructed Wikipedia article could send you scurrying off to investigate - but you wouldn't want that to necessarily be something in its favour!).

I think you are in a similar space to me at this point, and I'd love to hear your thoughts on the second and third parts of this enquiry (Tuesday 10th and 17th February).

All the best!

Chris.

[ptermx]

Hi Chris,

I thought I could sense that we were heading along similar paths.
I take your point about badly-constructed articles. Even worse could be sources that contain cunningly-constructed lies. Rather than try to be more discerning about which sources we allow into the assessment space, however, I'd prefer to remain focused on reader response and say that a valuable response is one that challenges the source or extends investigations into areas that it does not anticipate.
Of course, even then you may think of examples where this might appear desirable: where people are lured into apparently wasting much time or effort on trivial or intractable problems, for instance. But then, what objective standards of "trivial" or "intractable" are there? Whose time is it that was wasted? What would anyone else's opinion on that matter?
Anyway, I look forward to seeing your other articles in this series!

Peter

[Peter Crowther]

An interesting question: Does the complete edit and talk history of an article on Wikipedia, which ascribes provenance to each character and shows up actively contentious parts, gives you *more* information? Articles in a conventional encyclopaedia may not be credited and necessarily do not carry their edit history with them.

[Chris]

Thanks for your comments, Peters! I'll reply in turn.

---
Hi Ptermx,
Thanks for continuing our discussion! Aye, this is another danger of the Wikipedia, of course: that it is possible to willfully distort the information in ways that would not always be detectable. I'm not sure if this is worse than articles that draw erroneous conclusions from sketchy evidence though - many musical bands complain that their Wikipedia articles are bizarre, and it seems to be because interviews and press releases are used as a citation but then presented in terms of their implications. This usually stands because cross-moderation of the Wikipedia focuses on whether a claim has a reference - very little time is expended on checking that the referenced source does support the written claim! :)

I think the essence of your position isn't epistemic, per se - you are assigning a new kind of value to sources that is not about knowledge as usually construed. We could perhaps characterise this as 'incitement to know' i.e. a value about epistemic value. It's an interesting take on this issue, and I suspect it points to a kind of person - those of us who are compelled to know. :)

Hope you enjoy the next two pieces!

---
Hi Peter,
Nice to hear from you! This is a great question, and I think there is a strong claim that the edit and talk history does give you more information about the claims made in an article - but only when those claims are challenged. The information disclosed by talk is often about which claims are disputed - which is indeed something a conventional encyclopaedia is less efficient at uncovering! The majority of suspect claims within the Wikipedia are unnoticed, or undisputed by the cluster of nerds interested in that article, so I don't think this can entirely offset the problem.

Although this piece is written as an attack on the epistemic quality of the Wikipedia, the three pieces as a whole question the idea of an encyclopaedia as a source of knowledge at all, but only because they question the concept of propositional knowledge (more on this next week!). I think in the question of 'encyclopaedia vs wikipedia', what we're looking at is a difference in factual texture, if you will. For me, the texture of information in an encyclopaedia is radically different from the texture of information in the Wikipedia - to the extent that I can name cases when I would prefer one over the other. Which kind of makes it a shame that the Wikipedia is putting encyclopaedias out of business. :)

All the best,

Chris.

PS: did you get the invitation my wife sent?

[Peter Crowther]

I think you might appreciate the footnote at the end of http://www.theregister.co.uk/2015/02/08/worstall_i_support_the_green_party_natalie_bennett_minimum_basic_income/ in relation to this piece! I'll hold off further discussion until I've read the next article.

P.S. Yes, and we'll be there - thanks very much!

[Chris]

Hi Peter,
Sadly, I can't get that link to work, for some reason...

I'll look forward to hearing your thoughts as this short series progresses.
*waves*.

[ptermx]

Hello Peter. I too would be interested to see how/if/in what way exactly Tim Worstall might have found it within himself to "support" the green party, but the URL you supplied seems to be truncated and I couldn't find the article on vulture search....

[Dirk HK]

The correct link appears to be
http://www.theregister.co.uk/2015/02/08/worstall_i_support_the_green_party_natalie_bennett_minimum_basic_income/

As to the discussion: I agree that the edit history is a great means of discovering evidence to increase (or decrease) trust in the article. In particular because you can find whether some of the basic tenets have been challenged.
However, the absence of such challenges can mean that the argumentation is not "hardened" as it were, but also that it is self-evident.
Now, what appears to be "self-evident" or "obvious" may be rightfully questioned, as we all know, but this is true of all knowledge sources, encyclopedia, textbook or wikipedia.
I still fail to see why wikipedia is different here from classic encyclopedias, even though I agree that the "factual texture" is different. That the more loosely woven texture of wikipedia is necessarily a bad thing, this I don't understand.

[Dirk HK]

Ouch, link truncated again. Please merge the lines

www.theregister.co.uk/2015/02/08/
worstall_i_support_the_green_
party_natalie_bennett_minimum_basic_income/

[Chris]

Hi Dirk,
Thanks for the corrected link, and for sharing your thoughts on this discussion.

"I still fail to see why wikipedia is different here from classic encyclopedias, even though I agree that the "factual texture" is different. That the more loosely woven texture of wikipedia is necessarily a bad thing, this I don't understand."

The key difference between the Wikipedia and a classic encyclopaedia is that because you have limited information about who authored or edited a Wikipedia page (and about whether it has been sabotaged), you can never have *justified* true belief from such a source, even if the content there is indeed 'true belief'. To put it another way, the trust we can have in the process of assembling an article in an encyclopaedia exceeds that of any Wikipedia page. This argument rests on the 'justification' part of 'justified true belief'. (Of course, if you distrust experts, my latter claim might not be true for you! But then, the Wikipedia will be in no better a position... :D)

What might be tripping you up here is the implication that regular encyclopaedias are necessarily superior sources of knowledge. But I don't make this claim - I only assert that a conventional encyclopaedia is a superior source of *justification* for the claims it contains.

If you follow the link here to part two of this enquiry, "Factual Knowledge", you'll see that I'm knocking regular encyclopaedia's out of contention for being sources of knowledge as well! The whole notion of knowledge as disembodied facts is the target of this enquiry. The 'justified true belief' construal confuses us as to what knowledge might be. The Wikipedia is just my first casualty, and admittedly largely because I wanted a punchy opening. :D

Also, I think it is useful to discuss the epistemic values of the Wikipedia, especially since it claims to be an encyclopaedia, which I personally find to be a tenuous claim. I prefer the term 'trivia aggregator', personally - and it is the best of these we have! :)

All the best,

Chris.

[Chris]

Peter Crowther: *grins* That was worth waiting to get the link sorted out. :D

[Dirk HK]

I agree, Chris, with everything you said. And maybe the only disagreement we might have is that I am more content with having a good "trivia aggregator".
Well, ...
...no, there's another thing: For some issues (quite current ones), the trust you can place in Wikipedia by far exceeds the trust you should have in an encyclopedia. Take, for example, the former President of West Germany, Richard von Weizsäcker.
Even the English page (http://en.wikipedia.org/wiki/Richard_von_Weizs%C3%A4cker) gives his day of death. Consulting any encyclopedia (printed or digital) will not give you this for quite a time (unless electronic versions rely on a wikipedia-style updating mode).
Basically this boils down to the statement that you should carefully distribute your trust. (And obviously that "trust" is not boolean, but a gradient).

[Chris]

Dirk HK,
"And maybe the only disagreement we might have is that I am more content with having a good 'trivia aggregator'."

Not even this! I too value a good trivia aggregator: if I did not, I would not waste my time editing the Wikipedia! :) As I say in the concluding paragraph above, I 'appreciate its remarkable virtue as a public database.'

As for speed of updating - this is both a plus and a minus, of course! Certainly you can expect certain data to be revised far more rapidly in the Wikipedia. But this comes with the commensurate risk of it being sabotaged both more rapidly and more easily than a paper encyclopaedia. ;)

"Basically this boils down to the statement that you should carefully distribute your trust. (And obviously that 'trust' is not boolean, but a gradient)."

This is wise advice for any situation.
Thanks for continuing our discussion,

Chris.