I was thinking of the democracy-theoretic reason for the function here or on Reddit (the main examples I can think of at the moment), but then I remembered:

  1. When we vote for political officers, we tally upvotes only; it's not like the US president, for example, is decided by a ratio of upvotes to downvotes. Or, then, I've never heard of actual negative political voting (the notion is discussed in the linked SEP article, granted, but the Wikipedia article on disapproval voting seems to list only one obscure historical example in this vein, a temporary measure in the USSR).
  2. The thumbs-up/thumbs-down imagery on Reddit (or implicit elsewhere) hearkens back to Roman gladiator fights, doesn't it? Not the most inspiring tradition to base a practice on.

A further consideration: §2.3.1 of the SEP article on social-networking ethics reads:

... contemporary debates about social media’s alleged propagation of a stifling ‘cancel culture,’ which bend back upon the philosophical community itself (Weinberg 2020, Other Internet Resources), reflect growing anxieties among many that social networking environments primarily lack affordances for forgiveness and mercy, not judgment and personal accountability. Yet others see the emergent phenomenon of online collective judgment as performing a vital function of moral and political levelling, one in which social media enable the natural ethical consequences of an agent’s speech and acts to at last be imposed upon the powerful, not merely the vulnerable and marginalized.

Is that what happens on Reddit, or here on the SE, though? On Reddit, for example, QAnon was ostensibly driven out after their rhetoric (surprisingly rapidly) escalated to thinly veiled, or even outright open, incitement of violence. However, the cult didn't actually vanish from Redditspace but largely dispersed throughout a variety of subreddits, including (perhaps especially) the main one devoted to conspiracy theories. The ability to downvote QAnon lurkers when they rear their heads doesn't seem particularly effective at dissuading them from further propaganda efforts, much less at improving their attitude more generally. On the SE network, the saga of Monica Cellio (a rabbit hole I'd passed by many times before but that I delved into this morning) might or might not be directly pertinent to the ethical problem of downvoting, but insofar as her expulsion could be interpreted as a sort of negative meta-vote (by the SE administration), then it doesn't seem like the most pragmatic move to have made for the stated reason (I can't tell if she did anything genuinely malignant, for starters; and even if she did something wrong or worrisome, this fact was not made clear enough to onlookers and so negativity towards her arguably did not effectively serve the intended purpose).

So did "canceling" (or trying to do so) QAnon, or Monica Cellio, or whoever, "speak truth to power" effectively? Is this entire question too cutting-edge, so to speak, to be answered at this time, so that we'll have to wait and see about whether the moral quality of Internet discourse is further compromised, on average, for the better or worse, in connection with downvoting functions? I swear there were a few years, maybe before GamerGate, where it looked to me like the average was improving. I've been "terminally online" since my adolescence, so I've tried to keep up with these trends since then (when I was 10, my parents withdrew me from public school, moved to a remote island, and dissociated from all neighbors, church groups, and mostly even our relatives, so my socialization occurred almost entirely via the Internet), and that's what my "intuition" says (that there was a "grace period" when Internet discourse was positively weighted on balance), but I don't think I can trust my intuition too much in this case.

Clarification: one of Myspace's notorious features was the ranked friends list. I admit that it would unsettle me (or worse) to see someone I thought of as my best friend ranking me below others upon whom he relied less than he did me; but anyway, one of the reasons I gravitated towards Facebook "in the end" was the absence of such rankings. To boot, I appreciated that they had a Like function but nothing, at the time, in the way of Dislikes, and it has unsettled me all over again, although now indirectly, that Facebook shifted to having the ambivalent Laugh and Anger reactions available for posts. I have even wondered how much that change coincided with the seeming degeneration of Internet discourse, whether it was an effect or cause of this degeneration (probably both, but then to what extent either way?).

So the impact of social media on dopamine processes plays into my unsettlement overall. I surprised myself today to see that my upvote/downvote ratio here on the SE is roughly 1000-to-40; but then I know that I am wary of casting downvotes, not only due to the risk of a minor loss of my own score, but also because my mirror neurons are so entangled with my Internet sense that the thought of downvoting someone else makes me feel the kind of negative dopamine rush that they might feel were I to actually cast a downvote, and so it's hard for me to agree to express that kind of hostility online. (I'm slightly less ill-disposed towards Reddit downvoting, simply because of how easy it is to get a massive score there and anyway the score fluctuations are much closer to continuous than discrete in appearance; even so, I admit as well that any downvotes that I receive on Reddit weigh on me, regardless of whether I know they came from trolls or not.) Supposing that there is some kind of ethical responsibility involved in being able to influence other people's dopamine systems as such, then is the "weirdness" of downvoting, modulo the democratic ideal that theoretically allows for it, a function of the awkwardness some of us feel when it comes to such an influence? To my knowledge, one of the lawsuit situations Facebook faces (or could soon face) has to do with teenagers who have suffered various psychological/emotional damages in connection with social media usage. Is that danger(?) where my intuition of "moral weirdness" is coming from?

  • 13
    Negative political voting is ubiquitous. In referendums both Yes and No votes are cast. In many parliamentary systems (Spain, Israel, Australia) governments are installed and some measures are passed by relative majorities, i.e. votes are cast both for and against and the balance decides. Office elections often have 2-person runoffs where No is Yes for the other or the protest vote. Ranked voting is a graded equivalent of voting up and down. No is typically skipped for pragmatic, not moral, reasons.
    – Conifold
    Commented Aug 11, 2023 at 0:30
  • 5
    I would say that your squeamishness about downvoting, and your feelings from receiving any downvotes even if "you know that they are from trolls" simply marks you as a non-psychopathic and an empathetic person. Some people are not. I'm pretty sure that this characteristic is not taught to people and can't be. It is a watershed criterion. But, we are not supposed to hold such views, so I hope that someone can give a better explanation.
    – Scott Rowe
    Commented Aug 11, 2023 at 1:57
  • 11
    The point of downvotes on SE isn't to punish other users with negative meaningless ranking scores for displeasing you, it's to deselect incorrect answers. SE isn't a discussion site, it's a peer-moderated question and answer site. It isn't always successful (I've given a highly upvoted answer which I later discovered was wrong, for example), but it seems to do reasonably well. As far as I can tell the point of both upvotes and downvotes on reddit and youtube are the same as the point of reddit and youtube, which is just to increase aggregate time on site and thereby increase ad revenue.
    – g s
    Commented Aug 11, 2023 at 4:13
  • 5
    @ScottRowe I'm not an evolutionary psychologist and evolutionary psychology is barely science anyway, but I suspect that the downvote aversion is an evolved holdover from living in small primate tribes. Once you contextualize your aversion to the mild disapproval of one anonymous stranger out of seven billion as being based on the belief that the anonymous stranger is a seriously angered person you'll have to count on for survival for the next twenty years, the aversion goes away, no psychopathy required.
    – g s
    Commented Aug 11, 2023 at 4:29
  • 5
    I agree with gs that downvotes are there to indicate that a particular answer or question is bad, not that the user is bad. If anything, in some cases there is not enough downvoting on Philosophy SE. There are some atrociously wrong answers that don't get downvoted, and just hang around potentially misleading visitors to the site.
    – Bumble
    Commented Aug 11, 2023 at 7:57

5 Answers 5


Sharing trust judgements

Downvotes and upvotes are based on the idea of sharing information, and more specifically, on sharing whether some person or statement should be trusted or distrusted.

I'm no anthropologist, but I'd posit that this can be traced back to caveman times. I'd be very surprised if we don't see that in other animals (I don't mean cats are using Reddit or whatever, but rather the idea of sharing information and sharing trust judgements).

This, in its idealistic form, provides a useful signal to others to reduce harm that may come from misplaced trust, or time wasted that may come from establishing whether to trust someone or something for oneself.

Downvotes and upvotes are essentially quantifying the idea of "reputation" (in the social sense). If someone is a murderer, people say "that person is a murderer", and you know to not trust that person to not murder you. If someone lies, as one example, people may downvote them, and you'd know to not trust what that person says.

Quality control

Downvoting can also serve as a quality control moderation measure. If someone murders someone, and you say "no, murder is bad", that may tell that person and others to not murder people. If enough people collectively decide that murder is bad, and there's agreement that someone murdered someone, we may decide to put that person in a box for a while, as a consequence of having murdered someone. Similarly, if you downvote someone for what they say, that provides a signal to them and others not say what they said. If enough people collectively decide that saying certain things is bad, and there's agreement that someone said something (or lots of things, usually) in that category, we may decide to e.g. ban that person from a platform, as a consequence of having said that.

Affecting promotion

Downvotes also serve as a signal to a website, to (potentially) tell them to not promote something to others. This is on a similar note as trust and quality control, as discussed above.

So I think all of the above could serve make a case for virtue of downvotes (in theory).

Do downvotes actually work in practice?

Well, that's a whole other question, that probably has more to do with implementation, statistics and psychology. I certainly think there are a lot of issues with how they're implemented (on SE, if nowhere else), although I think removing them entirely may create more problems (possibly, maybe).

Downvotes making people feel bad

As for making people "feel bad", that is a downside, and you'd need to weigh that up against the potential upsides. This also has a lot to do with implementation: On SE, this directly and transparently affects your clearly-visible imaginary internet points, so that implementation is especially "good" at making people feel bad. On Reddit, this is much more behind-the-scenes and fuzzy, and post scores can't go below 0 (even if comment scores can, and even if you can get a good idea of score based on "percentage upvoted"), so I think that doesn't make people feel as bad (anecdotally speaking).

Making someone "feel bad" for doing something wrong can be argued is a good thing (if it serves to prevent someone from doing it again). But one has to consider (a) whether this is proportional to what they did wrong, (b) whether they know what they did wrong, (c) whether they even got the signal that they did something wrong at all, and (d) how effective this is at preventing them from doing similar things in future.

On certain sites (that e.g. are more just for fun), there might not be much moral good or harm that comes from what people say, so making people feel bad with downvotes might not have the same moral upsides. On the other hand, downvoting helps to keep things on topic, which keeps the site fun for others, no-one is forced to use such sites, and you enter into an informal contract that includes downvotes by using the site. On the other other hand, people don't explicitly consent to downvotes and that might not need to be part of said contract.

In conclusion

There's no clear conclusion about whether downvotes are right or wrong in general. It depends a lot on the specifics and implementation and how important you consider different factors to be.

On some more specific notes:

  • Downvoting QAnon (or any other movement) could help to limit their effectiveness at spreading their message on a particular platform.

  • Monica wasn't really downvoted, to the best of my knowledge. That was more just related to reputation in the classic sense (someone saying something bad about someone, which I suppose you can call a "meta-vote" if you want), as well as SE deciding to end their working relationship (for bad reasons), which is distinct from sharing trust judgements, etc. that downvoting is used for.

  • Mentioning Roman gladiator fights sounds a lot like (fallaciously) using something's origin to try to invalidate it. Also, on this site, we use triangles for upvotes and downvotes (but triangles are mini-pyramids and pyramids schemes are bad). So this is arguably a UI objection rather than a moral objection to the concept of downvotes.

  • I'd argue that most of the problems with internet discourse (and, by extension, many problems in the "real" world) stem from how social media was designed, that anyone's words can spread to millions of others all across the world in a matter of hours or minutes (regardless of whether or not those people want to hear what that person has to say), that it encourages constant engagement, that they might see negative engagement (e.g. downvotes) as a positive signal, or as a neutral signal (this isn't the case for downvotes on SE, as an example), etc.

  • Clear, complete answer that helps me think about the issues. When I was a kid, and some other kid was acting up despite lots of negative feedback, adults would say, "he's just looking for attention." Not sure if they still use that phrase. I didn't really want attention from adults, so negative attention was definitely out for me.
    – Scott Rowe
    Commented Aug 11, 2023 at 11:49
  • I think this answer is helpful because it helps me understand my concerns on a deeper level. For example, I don't believe in the concept of people being trustworthy/untrustworthy and am broadly skeptical of character trait theory, and I'm also nonplussed by the (in)efficiency of criticism in changing behavior. Regardless of whether those of my beliefs/disbeliefs are justified, it's good for me to recognize how they play into the issue I'm having. Commented Aug 11, 2023 at 12:41
  • 1
    at my school, i don't remember any children being disciplined (they called them "minuses", like a warning about detentions) for doing poor work, rather than misbehaviour or just not doing the work
    – user67155
    Commented Aug 11, 2023 at 16:33

If 'moral' means something like:

"Concerned with the principles of right and wrong behaviour",

and if 'weird' means something like:

"very strange and unusual", then

It's perhaps not too much of a stretch to deem 'morally weird' as almost tautological, in the sense that the diversity of moral opinion ensures that an individual's morality will likely seem "very strange and unusual" to at least some significant proportion of the population. Similarly, a downvote that seems strange to one person may seem perfectly appropriate to another.

Downvotes are perhaps best utilised as warnings to others that an opinion might not be as reliable as we'd like; that an answer is suspect; at odds with evidence.

Downvotes are perhaps worst utilised as punishments for 'wrong' opinions or as disparagements of perceived personalities/belief sets.

There is nothing weird about this. We are dealing here in a forum of ego and opinion and power. We know that when we are granted power over others, we are frequently tempted to use it maliciously. This knowledge should help us to recognise that downvotes do not necessarily reflect true or substantial flaws in our contributions. It is sometimes tempting to give too great a credence to criticism, and to reflexively doubt our opinions merely because someone has disagreed with us.

However, we should always keep in mind that downvotes which seem at first unreasonable may after greater consideration reveal flaws in our thinking to which we'd previously remained oblivious. Whenever we feel the pang of resentment at being downvoted, we should bear in mind that we might in fact be the beneficiary of insight into our own shortcomings.

So. How to differentiate between 'good' downvotes and 'bad'? Consistently apply - to the best of your ability - the rigour of logic and evidence to your questions and answers, so that when faced with criticism (and compliment), you can reliably assess the merits of your contributions and the merits of you detractors.


In an election, you're making a choice between candidates. So an upvote for one candidate is effectively a downvote for all the others. There's no need to actively downvote each of them. There are also systems like ranked-choice voting, where you do indicate how much you "like" each candidate; but it's still not necessary to express dislike explicitly, just leave them at the bottom of the ranking.

But the "like" and "unlike" buttons on social media are not for choosing between alternatives. These are quality judgements, and it makes sense for people to express "this is bad". Each item stands on its own; when you dislike something, you don't have to have an alternative that you prefer instead of it.

In effect the alternative is "everything else that I haven't disliked", but that's so broad that it's not very useful. If you say "I don't like to eat broccoli" you don't have to state which vegetables you would eat instead.

  • But see that binary "if you're not for me, you're against me" standpoint seems inconsistent with the distinction between "believing that not A" and "not believing that A," and then with the intuitionistic critique of proof-by-contradiction. OTOH, I tried out a "nonnegativity" standpoint for a while, but quickly realized that the prefix "non-" in play meant that my attitude undermined itself by default. Commented Aug 11, 2023 at 14:39
  • 1
    @KristianBerry True, since there's also the option of abstain, which could be considered equivalent to "I don't believe that any of the candidates is fit for office."
    – Barmar
    Commented Aug 11, 2023 at 14:54

One issue about social media, life in general, is how we reproduce the illusion of community, control, freedom, knowledge etc.. Obviously, many sorts of condoned behaviours - especially on-line - can be vicous and vice forming, and yeah I agree that there is a psychological tendency toward bad habits in down, more than there is up, voting.

There is something suspect about it (sure, not as bad as the thrill of up-voting a politically dangerous meme), analogous to shushing someone when everyone else is shouting. So I don't think it's wrong, even if how it is presented may be, but not exactly an ideal.

  • 2
    It can certainly be much worse than ideal. For instance, someone posts an imperfect question on stackexchange, and immediately seven people gang up to downvote and close it. This would be the equivalent of a newcomer entering a room, timidly asking a question, and seven persons in the room jumping at them to tell them how utterly wrong they are. Such behaviour would be called bullying in real life, and wouldn't be acceptable, but it's commonplace and even encouraged on stackexchange, because downvoting a bad question is considered a good thing.
    – Stef
    Commented Aug 11, 2023 at 16:45
  • i don't know how you define 'bullying', but it is certainly the case that bullies often @Stef believe they are right (that their target is e.g. weak)
    – user67155
    Commented Aug 11, 2023 at 16:47

Is there something "morally weird" about social media sites with downvote functions?

Yes. I don’t have any social media accounts. But PSE allows downvoting anonymously and without explanation, thus freeing the philosopher from responsibility for their actions.

This rule is weird because PSE is supposed to be a site where people can express themselves freely on philosophical matters. In exchange for a moderated site and the ability to pick an anonymous online handle, people should take responsibility for their negative evaluations.

When I write "express themselves freely," I mean that few philosophical questions are out of bounds on this site. And fewer still are so out of bounds that simply asking the question should prompt a downvote.

  • You've not explained why it's "morally weird", other than the lack of explanation.
    – Dragon
    Commented Aug 12, 2023 at 10:29
  • 1
    @Dragon. How’s that? Commented Aug 12, 2023 at 18:15
  • 1
    SE is not to express oneself freely on matters, but to ask for and get provided credible expert answers that should be exhaustive and sourced. It is a main problem of this particular community (philosophy.SE) that both answers and votes are more often a free expression of opinion rather than being based on actual expertise.
    – Philip Klöcking
    Commented Aug 12, 2023 at 18:34
  • @PhilipKlöcking maybe we can look at the situation as relative to a period of time more than as characteristic of this SE in itself. Conifold and Bumble have very high scores mostly because they know the source material per the SE's protocols and post questions and answers (over a "long" period of time, granted) on that basis. The same goes for many other posters with the "very high" upvote scores. Now sometimes, like now it seems, it might seem like some newer users have unusually high scores for the quality of their citations, but what is the direct cause of this? Commented Aug 12, 2023 at 19:00
  • Ironically, some users might have these unusual above-average scores for reasons of some background pressure to offset the range of the downvote score. Like, some might "cheat" and have multiple profiles, or some might express an opinion popular with a set of lurkers who happen to be here on that day (or whenever), and so maybe even that would be a "motive" for someone to post seemingly excessive numbers of questions in the same day/over a few days. Etc.? I wonder if it would help to make downvotes as impactful on score as upvotes (and if there's any 1:1 dopamine-level correlation). Commented Aug 12, 2023 at 19:04

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .