The Myth of the Season

I just started playing Resonance of Fate. Great game so far.

Fans and critics of anime across English-speaking communities over the internet have all been operating under a terrible misunderstanding. It is a misunderstanding that, I believe, will require significant effort to correct on the part of hundreds of people, and will take much time for us to overcome – particularly for the blogging and fansubbing communities. What could such a glorious misunderstanding be?

What I am referring to is none other than the familiar term “season.” The liberal use of the word “season” as a replacement for the phrase “anime airing for a set duration of the year” has led us (bloggers, fansubbers, and fans alike) to believe that there is such a thing as an “anime season.” The usage of the term “season” is, in my opinion, a scourge upon methodology. And methodology is crucial if we want to study anime in the long term, which is the topic most often at hand when we invoke the term “season.”

First and foremost, I want to emphasize that this problem strictly applies to blogging about trends in consumer behavior and production of anime. I am not arguing that the term “season” should be eliminated all together; what I’m calling for is an acknowledgement of its perhaps dubious methodological status. That being said, it becomes relatively clear why methodology is so crucial to our understanding of trends in anime production over the years: we employ a statistical analysis with various amounts of rigor to discuss trends. But the most critical part of this whole project to analyze trends is that we don’t place the same standards on our statistics.

People are, in essence, taking for granted that their conscious and subjective picks are representative of and constitute as evidence of some sort of idea that they have in their minds. It’s like taking correlation for causation, something that we’ve all been taught not to do. Except there’s no correlation; it’s 100% epistemically unfounded because of its subjectivity.

Let me provide a sample of what I mean. Looking back on last month’s posts on this blog alone, four particular posts stand out to me as referring to an “anime season.” Sorrow-kun wrote a two-post discussion of anime that aired during Winter of 2010, in which he discusses two particular series, Sora no Woto and Durarara!!, that have been memorable, and argues that another two enjoyable series should be excluded from an analysis of “season” on principle of not being influential, these being Hidamari Sketch Hoshimittsu and Nodame Cantabile ~Finale~. In comparison, let’s look at the number of shows that aired that Winter: 22 weekly shows and 12 new OVAs. Intuitively, it seems quite strange to consider this an analysis of a season when it covers so few shows.

Now, some of you may argue that some of these shows are continuations, and that OVAs also shouldn’t count. It’s like how the anime blogosphere, when discussing “seasons,” arbitrarily chooses to neglect long-running anime from their analysis (One Piece, Naruto, Bleach) which may be simultaneously airing as well. But I think that this merely proves the arbitrariness by which we use the term “season.” If a season is a set period of time, and not some figment of imagination invented by bloggers to discuss how good or bad some arbitrary set of hand-picked anime is, then, by methodology, we should not be excluding continuations.

I want to write something about Haibane Renmei next.

On the other hand, Akira has a post discussing a possible change in the moe trend we’ve seen in recent years by taking a look at Spring 2010. The difficulty with using the term “season” here is that it’s obviously far too early to make claims about both producers and consumers disliking a moe-centric model of anime design. Although Akira acknowledges that it’s too quick to judge, I think the more crucial problem is whether or not we can use something as dubious as a “season” to discuss a change in a trend that has lasted for decades. After all, moe has been around for a very long time. To discuss a the decline of something that has lasted for tens of years in one incomplete “season” seems to me to be comparing apples to oranges. Perhaps that’s why he added a question mark to the post’s title.

There are several reasons why this anomaly of a term came to being. Here’s the first reason why the term “season” is deceptive: when we talk about T.V. or radio, we don’t use the word “season” to discuss the same thing we do when we talk anime. A “season” for a T.V. show is a number of episodes that are produced and aired in succession, and generally tend to cover a single arc in a story (more or less). Thus, it’s clear that this use of the word “season” doesn’t invoke a set period of time in which multiple shows air. Then why do we use it for anime? The answer, I think, lies in confusion between the four seasons and the use of seasons in broadcasting.

The second reason why this is a problem is simply the fact that subjective choice over which anime to discuss results in overgeneralizations. Part of the problem is that people use the term “season” mistakenly in the first place, overgeneralizing about all the anime airing in that “season.” The result is an attempt to justify contradictory feelings about the anime that had aired, and thus results in more overgeneralizations. What we want to do is vindicate our overall impressions of several anime that saliently stand out to us, and this causes us to group everything into a collective term “season.”

I am not, in any way, trying to pick a fight with my colleagues, whom I respect greatly and have much affection for. However, I am proposing a challenge to those who think that the term “season” should be used. Demonstrate to me how “season” can be particularly useful as a unit of analysis without having dangerous pitfalls. If this can’t be done, I believe we should make a few changes. First, we only make attributions of “good” or “bad” to single series, and not to entire “seasons.” Second, we analyze trends in production and consumption using a unit based on a much longer time span. What I have in mind is something like what aquabluesweater does over at Blue Sweater Story. The data is clean, neat, and discusses trends easily. Granted, some arbitrariness will always be involved (since multiple animes air across any cutoffs we decide to impose). However, by expanding our analyses, we may find ourselves more accurate and methodological in our critiques, and less inclined to generalize about series in general.

12 Responses to “The Myth of the Season”

  1. Huh. While I don’t disagree with you, I think you misunderstand something.

    There are several reasons why this anomaly of a term came to being. Here’s the first reason why the term “season” is deceptive: when we talk about T.V. or radio, we don’t use the word “season” to discuss the same thing we do when we talk anime. A “season” for a T.V. show is a number of episodes that are produced and aired in succession, and generally tend to cover a single arc in a story (more or less). Thus, it’s clear that this use of the word “season” doesn’t invoke a set period of time in which multiple shows air. Then why do we use it for anime? The answer, I think, lies in confusion between the four seasons and the use of seasons in broadcasting.

    It’s pretty common that people refer to season of anime in that precise term, which is exactly why the term is not used to describe long-running shows. Far majority of the time the term season is used to describe, as you said, a number of episodes that are produced and aired in succession. Usually that is the entirety of all 1-cour shows, and it wouldn’t be wrong to refer to that as one season.

    The term is confounded when shows tend to air in succession, weekly, when you have 4 cours a year, as it coincides with the climatic seasons. But I don’t think most of us confuse this, just some of us.

  2. I think the problem is less with the word season itself and more with the methodology as a whole. The blogging format is, after all, a rather informal one, and as such people are willing to speak colloquially and skip the rigors of thorough explanation. We could drop season for the more stringent term cour (or at least I would consider it a stricter term), but I believe the same problem would arise.

    Sorrow-kun’s post intrinsically implies, or at least anyone who has a notion of critical thinking should be able to pick out, that he’s bringing up his own opinion on select shows that are airing across this cour that he has seen that have interested him in one form or another and which he believes is to some extent representative of the shows that began during this cour. If you think that’s a mouthful, that’s because it is. There are many things his post must be grounded in first in order to have even written such a post. One could make an argument over particular assumptions he made, such as whether or not the sample he watched was accurate in its representation of the season or if he can so easily discard the bad samples when judging a cour, but that’s where the problem lies, not in the word season. It wouldn’t matter what it’s called, as the the problem arises anyways. He simply needs to set his parameters more rigorously.

    Likewise, in Akira’s post, the observation is again grounded on some key assumptions. If he really wanted to be clear, he could perhaps say “There has been a such and such trend in moe before, but a certain set of series during this cour that begins between April 1st and April 31st seem to buck this trend.” Again, there are some fallacious arguments here, such as whether the set of series is a representative sample of the entire season and if the single cour can show any kind of change in trend or if it’s simply an outlier. However, this comes from deriving a sweeping conclusion from a narrow sample, the boundaries of a season being an irrelevant point even after redefining with clearer terms.

    It would, of course, be nice to have a strict definition for season and long units of measure are relevant when, as you said, it comes to gauging trends because of the passage of time (although I like to consider cours to be the smallest unit of time to use due to that being the shortest length of a single series, followed by probably years), but I do believe that cleaning up the methodology itself would solidify the analysis of trends far better than providing a word that gives us a set unit of time. The problem of using seasons as a measure of long trends is in itself a methodological error instead of a misinterpretation of the vague definition. For myself, I’m fine with people simply bringing up observations or conclusions in an informal manner, because that’s kind of how blogs are: opinionated and personal. Such things are certainly noteworthy and should be taken into consideration when discussing trends as long as one can evaluate the limits and the subjectivity of such evidence or conclusions. We don’t really need to to hit a peer review level of rigor to discuss such things.

    I now expect to go orz when Kylaran comes around and tells me I completely missed the point.

    And I hope that Haibane Renmei piece turns out nicely. I haven’t seen a commentary on that in a long time and it holds a special place in my heart for being one of the first series to get me into anime :)

  3. Every idea has “dangerous pitfalls.” I think the idea of a season makes more sense than most. A season isn’t so much a unit of time as a unit of competition among anime producers: they all produce animes to come out on TV during a particular 13-week period, and to be sold as DVDs then and subsequently. They are competing for the money and attention that are spent on anime during that period. The nature of the animes is an indicator of what the producers think consumers want to watch or buy — as well as of what producers and creators want to make. TV ratings, comments, and sales show what consumers actually want.

    Shows that go two seasons (“runs” or kuuru, in Japanese) can be used as an indicator over both. Long-running shows like the ones you mention are the underpinnings of the whole system, showing what kinds of animes remain popular through the years. And, of course, the shows with by far the highest TV ratings are those that most bloggers say nothing about, since they are “family” shows for young children. But that is not the audience whose evolution bloggers are writing about.

    Anyway, that’s how it seems to me.

  4. Here’s the first reason why the term “season” is deceptive: when we talk about T.V. or radio, we don’t use the word “season” to discuss the same thing we do when we talk anime. A “season” for a T.V. show is a number of episodes that are produced and aired in succession, and generally tend to cover a single arc in a story (more or less). Thus, it’s clear that this use of the word “season” doesn’t invoke a set period of time in which multiple shows air.

    But that IS how TV is talked about. The fall season. Mid-season replacements. The summer season (big for cable shows). The only real difference is that “season” can mean both the year number as well as the time period, and they don’t always match up (although some times they do, like with mid-season replacements and cable shows.)

  5. Thanks everyone for the replies!


    I can see the reason why people use the term “season” in the way I’ve outlined above; however, what I want to point out is that there is no real reason to use this term when supporting our claims. Can we really call a season good or not? Is there any point in arguing that a bunch of anime airing Fall of some year is better than the anime that aired Spring of another? I just find that these comparisons are far too arbitrary. There’s nothing wrong with season being used outside of this context.


    Don’t worry, you’re right on the mark. :) I think the problem with my post is that I’m not explicit about what sense I think the term “season” is being misused. It is mainly the methodological sense, which is what you point out.

    I think you’re right in that opinions on seasonal anime are worth reading. But I don’t think we can rely on seasons to compare trends or compare between “seasons.” It’s perfectly fine to point things out — similarities, differences, etc. — but a stricter comparison should be made. Cours would be a much more satisfactory term here. But I don’t expect bloggers to do THAT much research; we’re not, as you said, the peer review board for some academic journal.


    You’re right. There’s perfect sense in using the season or cour as the standard from a business sense. However, if we’re going to attribute good seasons, bad seasons, changes in trends as a new season’s shows start airing, then we should be much more careful of using terminology that I don’t think was designed for the purpose of critiquing in the way that people like to about the types of shows and consumer demands.


    Right, and the use of seasons should continue to be used that way. The more problematic thing here is that “season” is being extended beyond this context and I just don’t think there’s much meaning using “season” from a comparative standpoint. We should be comparing individual anime, not seasons, and long term changes. The idea of the “season” being a unit for comparison just seems to me to be far too murky to employ carefully.

  6. I think the problem is less about “seasons” being a bad measurement and more to do with analysis which is often at best uncomprehensive and at worst willfully specious. I mean, American TV seasons for example have pretty distinct aspects to them (fall for dramas/sitcoms, spring for reality, summer for cable dramas, etc.), and there are noticeable trends in anime uh, seasonal seasons such as (to be incredibly obvious) that spring has more premieres than summer, or that big titles are rarely launched in winter.

    What I don’t know is exactly why anime seasons are broken out the way that they are. American TV seasons are centered around sweeps, sure, but that’s not applicable to anime.

  7. That’s an interesting question regarding the ‘quality’ of each season or cours. That would be a fun set of analysis for me to do. The most ‘scientific’ way I can see this being done is to obtain the total number of shows for each season and apply each to a rating (Anime News Network rating, Japanese TV rating or any other source where people actually review and rate a lot of shows). We can then weigh the scores appropriately and rank order each season. I can probably use the data that I have collected (based on all the top animes of the decade lists on various blogs) to measure the quality of anime seasons pre-2009 as well.

    On the moe front, if I have the data above, I can always mark all the moe shows and see over the long term the trend of the number of shows as % of total produced.

    Will let you know when I get around to doing all of these…

    PS. Looking forward to reading your article about Haibane Renmei and thanks for a mention of my blog:)

  8. Kylaran, correct me if I’m wrong, but I think you’re more against sweeping generalizations and trend-analysis rather than that a “season” isn’t a well-defined unit of time. I can see your argument made for any generalization of a period of time.

    I think you’re right in saying that we cannot objectively judge a season as being “good or bad.” But I go on this blog to listen to other people’s opinions, so I see no problem with all of you expressing yourselves.

    Most of this anime-trend analysis is just hopeful speculation. But I see no reason to be precise and scientific about such a subjective analysis. I mean come on, this is a blog not a scientific journal.

    @aquabluesweater I do think it would be an interesting article to read about a numerical analysis of actual anime trends! It would be interesting to see some actual numbers for once to put these opinions to the test. I am one of the few who believe that the big-bad-moe-monster isn’t as horrid as some may think.

    @everyone: I honestly really like the use of “seasons.” Yes, they are inaccurate; but I think its a lot easier to keep track of new anime if we put them into categories. I like the fact that I can prepare myself to watch a certain amount of anime in the “summer season,” even if the “summer season” is nothing more than a mere abstraction.

    I think its unfortunate that I only hear trend-analysis in January when the year is over. Sure, the amount of shows in a season is few, but that lack of accuracy is what make the predictions so fun! 😛

  9. I know this article is about stuff I wrote, but I really don’t have all that much to add. Yes, it’s dangerous to use “season” as a unit of analysis, and yes, a single season is far too short to make definitive comments about trends with really long time constants. But, personally, I wasn’t so much trying to do an analysis as I was posting a few impressions based on opinion and my own limited exposure to what was on offer during Winter. The fact of the matter is, it’s impractical for a single person to do a thorough analysis of a season. At the least, you need to take a cross-section of opinions from a wide array of people to cover the gaps. But I think we all knew this already.

    Anyway, I don’t think the definition of the word “season” is in dispute here, but then again, neither is the inability of a single person to make a comprehensive assessment of the quality and impact of a season and the direction of trends based on seeing a fraction of the titles on offer.

  10. Can we really call a season good or not? Is there any point in arguing that a bunch of anime airing Fall of some year is better than the anime that aired Spring of another?

    Yes, and yes. It’s similar to what JP said, that the problem arises more along the lines of the quality of analysis than if there was a legitimate reason for the mode of analysis.

    I think, to temper things somewhat, there are some overriding contexts that westerners miss when talking about TV anime, which might make more sense than a concurrent-batch quarterly (to avoid the s word) analysis. After all, businesses do quarterly analysis all the time, and there are very good reasons to do the same as fans in this particular case.

  11. Pardon me for taking so long to reply; I dealt with a barrage of midterms and papers this week.


    When analyzing what shows are aired and for what reasons, we should employ the term “season.” Even if the analysis is questionable, there’s a certain necessity to use season in this sense. On the other hand, there’s no real need to use the term “season” if we’re generalizing about how good a set of concurrently airing anime is from a small set of particularly salient shows that we happen to watch. We might as well be comparing individual anime, and dispensing with the illusion that we’re comparing seasons at all in this case.


    Looking forward to your work. Viewer ratings are a great way to compare seasons, assuming the relevant statistics is applied. Also, I appreciate your work a lot because I deal with data analysis all the time at the lab I work in.

    @blue cheeze

    First of all, your example is not counter to my point. I think it’s perfectly fine to use season in that sense. You’re psyched about anime that are airing at a concurrent time. Hence the use of season.

    Let me give you a clearer example of what I mean. When someone says “This season is bad compared to the previous season, and let me show you why through x, y, and z anime,” I think the term season fails to really pick out anything in the real world.

    But, if they wrote “I don’t like anime from this season because I don’t think they’re as good as anime from the previous season,” they wouldn’t be employing the word season incorrectly.

    I have nothing against expressing one’s opinion. I think we should just clean up our usage of terms a bit.


    Bingo. The word “season” isn’t the problem here — it’s the usage of it. There’s no problem with not being able to comprehensively take a look at an entire season; just drop the term season from your analysis, and you’re good to go.


    When businesses talk about seasons, they’re analyzing it not from “I think 3 anime are good because the plots are amazing, so this is a good season.” They look at it from objective data, viewer ratings, etc. etc. I’m all for calling seasons “good” if the analysis is business-related, but if we’re simply describing how much we like the shows we watch one season from the shows we watched another season, we should just drop the “good/bad season” business.

    Let’s look at this in mathematical logic. For (x)(y)(F), [F(x) = aF(y)]. If a season is good, then all the anime airing in that season is good. This is implied in the definition of season in the seasonal sense. But when we’re talking about only a select few anime from which we generalize the term “good,” then we should just drop the term “season” and say we’re talking about a select set of specific anime.

  12. […] remember making a comment a while back at one of Behind the Nihon Review posts about doing a numerical analysis on whether moe anime series has become more prevalent in recent […]

Leave a Reply

Gravatar enabled.