on “category errors”

From a chapter of my book called "Five Paths to Confusion," which discusses common errors in thinking about transit.   I hope that somewhere out there is a logician or linguist or philosopher who can suggest a better term for "category error."

Spectrum for blog

A category error occurs when you think about a spectrum as though it were a series of boxlike categories.  If you and a friend disagree about whether something is blue or purple, you’re making this error together.  Blue and purple are adjacent zones on a continuous spectrum of colors (technically wavelengths of light) and zones on a spectrum can only have fuzzy or arbitrary edges.  So if you disagree about whether something is blue or purple, you can both be right, based on slightly different notions of where you mark the boundary in the fuzzy area where blue shades into purple.  If one of you is right and the other wrong, it can only be because of some arbitrary standard about where blue ends and purple begins, a standard you’ve both agreed to respect.

Category errors are built into the very structure of our language.  Our category words feel like boxes with hard edges:  blue, purple, tall, wealthy.  But like colors, most of them really refer to directions or zones on a continuous spectrum.  There’s no objective basis for saying “Jim is tall,” unless we just mean “Jim is taller than most people.”  “Tall” is not a box; it’s just a range or direction on a spectrum of possible heights.  We all know that, and for simple ideas like height or color the category error rarely causes trouble.

But when we talk about emotive categories, such as wealth or success, we can easily lose sight of the spectrum, and as with blue and purple, this can cause pointless arguments.  Consider a famous comment widely (if falsely) attributed to British Prime Minister Margaret Thatcher:  “A man who, beyond the age of 26, finds himself on a bus can count himself as a failure.”  Stated that way, “a failure” sounds like a box that you’re either in or out of.  Associating buses with failure or poverty is a common attitude in some cities.  If you think about failure or poverty as a box, this can be an easy way to decide that buses aren’t worth your attention, and that there’s no point in thinking about how buses and rail transit can work together as one network.

But even if it’s true that bus riders are poorer than rail riders on average, you can change your perspective by reminding yourself that the boundaries of “poor” and “middle class” and “wealthy” are as fuzzy or arbitrary as the boundary between blue and purple.

Some category errors are built into transportation planning jargon.  For example, you may hear certain planners divide transit riders into two boxes.  One box, called a discretionary or choice rider, contains people who have the option of driving, and who will use transit only if it out-competes their car.  In the other box is the transit dependent or captive rider, who has no viable alternative and therefore has to use transit.   Dividing up riders this way leads to the idea that transit must compete for choice riders, while captive riders can largely be taken for granted; they will ride no matter how poor the service gets.  

These categories are imposed on reality, not derived from it.  Transit dependence, like wealth itself, is a spectrum, with vast numbers of people in the grey areas between “choice” and “captive.”  For example, many people with low incomes own a car out of necessity, but experience owning a car as a financial burden.  Many low-income families feel they have no alternative but to own a car for every adult in the house.   If we give these people credible alternatives to car ownership, they can experience the result as liberating, even though some transportation planners will now call them captives.   Often they will find better things to spend that money on, such as education.  Many people are in situations like these, and we can achieve both environmental and social goods by helping them choose to own fewer cars. The two-box model of society, where everyone is either choice or captive, prevents us from seeing those possibilities.  

Very few of our category words describe things that are really boxlike categories.  So whenever you hear someone divide people into two categories, it's worth asking, "are these really boxes, or just zones or directions on a spectrum?"

27 Responses to on “category errors”

  1. EngineerScotty March 23, 2011 at 8:57 pm #

    Obviously this is just an excerpt, but probably one of the biggest category errors in transit is the concept of “rapid”.
    And for the example you cite–whether or not a patron is a “choice” patron or not, economists already have a way of describing this phenomenon formally–the concept of elasticity (in this case, of demand). A so-called “captive” rider is one whose demand is relatively inelastic; a trait exhibited by “hostile” riders as well (ones who refuse to ride transit under any circumstances). So-called “choice” riders have elastic demands for transit; and where there is a large number of consumers for some good with a highly elastic demand for that good–you run into the phenomenon of induced demand–simply supplying more will often increase consumption.
    I can’t wait for the book to make its appearance!

  2. Annonymous March 23, 2011 at 9:06 pm #

    The frequent categorization that people who own cars are “choice” riders and people who don’t are captive riders is completely bogus. It totally ignores the fact that even for non car-owners, transit still has to compete with other modes of travel, for example:
    – walking
    – running
    – biking
    – skateboard/scooter/segway/roller-skating
    – carpooling with family/friends
    – ridesharing on craigslist or similar service
    – vanpools
    – other agency’s transit services (Amtrak, ferries, etc.)
    – asking friends or relatives with a car to pick you up
    – courtesy shuttles (yes, a good transit system can often out-perform them)
    – private shared-ride services (e.g. Super Shuttle)
    – riding a taxi
    – renting a car
    – avoiding the trip altogether (for example, shopping online, rather than in person)
    And, last, but not least, buying a car (being able to afford a car is not black and white – the worse the alternatives, the greater the opportunity cost one is willing to forgo with a car purchase).
    Great post!

  3. Rob March 23, 2011 at 9:40 pm #

    I actually think the “choice” and captive” rider concept is more useful as a method for evaluating service, and the mindset of transit planners and officials, than for categorizing transit markets. There are clearly transit services that everyone knows will only be used by people who have no choice. But better transit systems are planned by people who believe everyone has a choice, and for whom the concept of choice vs. captive riders is irrelevant.

  4. Heather March 23, 2011 at 9:44 pm #

    I’m curious as to why you believe the Thatcher quote falsely attributed. Google sadly turns up nothing definitive either way. Although it is referenced in the Hansard. http://www.parliament.the-stationery-office.co.uk/pa/cm200203/cmhansrd/vo030702/debtext/30702-10.htm
    That aside, excellent post! My city is currently in the throws of a rail vs. bus debate and there is a vocal group that firmly believes the rapid transit proposal is waste of money, since only the poor and student use transit. One went so far as to argue that provide all lower income residents a smart car would be cheaper than building better transit.

  5. Tom March 24, 2011 at 1:16 am #

    Hi Jarrett,
    Many thanks for the wonderful and informative blog. I have thoroughly enjoyed reading it.
    As for your question about a philosophical name for what you describe under the heading of category error, I would suggest “over-determination”. Over-determined categories are too strongly determined, i.e. they are fundamentally illegitimate, or as you say of blue and purple ‘arbitrary’.
    I should mention that philosophy is a many-headed beast, and this precise sense of ‘over-determined’ is not universally shared. Over-determination here relates most closely to the work of a French philosopher named Jacques Derrida, who became famous (if controversial) for his concept of ‘deconstruction’.
    He analysed what he called ‘unstable binary oppositions’ (blue and not-blue as opposed to blue and purple but the principle still holds). For him, each term determines but also unsettles its opposite due to a hidden and structural preference for one term over the other; presence over absence, certain over uncertain, good over bad, (and train over bus?). His insight was that the complexity of the discarded term also needed to be understood in order to say anything meaningful about the preferred. However, we have learned a system of categories (the “metaphysics of presence”) which thoroughly conceals this from us. The effect is that categories seem ‘natural’ to us. The work of deconstruction is then to denaturalize what seems natural in order to reveal the investments we have in a particular (i.e. categorical) way of looking at the world.
    All the best with the book!
    Tom

  6. ant6n March 24, 2011 at 5:54 am #

    @Heather
    I believe in rail, and have made the same argument myself in the past (that one should just buy car for every user instead of building rail). The problem is that some rail is ridiculously expensive, compared to how many riders it will serve. When you get into the category 😉 of something like 100K$ in capital costs per resulting weekday rider, then you should probably start asking whether the project couldn’t attract more riders at a similar cost, or whether the cost could be reduced.

  7. John March 24, 2011 at 5:56 am #

    Margaret Thatcher: How old are you?
    Me: 31
    MT: Why are you on this bus?
    Me: I’m transit-dependent.
    MT: You must be poor. You’re a failure.
    Me: 🙁
    Jarrett Walker: Don’t worry. You’re probably poor, but only comparatively on an arbitrary scale.
    Me: Thanks. I feel better now.

  8. Tom West March 24, 2011 at 7:15 am #

    Here’s a scenario: a household owns one car, and has two people who wish to commute to totally different locations. Therefore one of them has to take transit, and one can drive. Is the person taking transit a “captive” or a “choice” rider.
    Another scenario: a person lives somewhere taking transit is substantially faster than driving ever could be (e.g. rail services to London, UK). If that person owns a car but commutes by rail, are they a “captive” or a “choice” rider?
    My point not to get answers to these questions, but point out that even something as seemingly black-and-white as “captive” vs. “choice” has some grey areas.
    (Also, one can be rich and not own a car,so transit-dependency does not imply low income)

  9. Alon Levy March 24, 2011 at 12:04 pm #

    I think I agree with all the above comments. I’d add that good ridership models are careful to avoid category errors. If transit is 20% slower than driving, they won’t say transit gets a 0% mode share – just a much lower mode share than if transit is 20% faster. See for example SNCF’s proposals for US HSR.
    Two things are happening here. First, different people have different circumstances. Maybe driving is 20% faster, but finding parking is difficult, and that raises travel time. Maybe I’m carrying bags, so that transfers are exceptionally inconvenient. And second, people have different personal preferences – for mode, for willingness to transfer, for tolerance of traffic jams, and so on. For the exact same intercity trip, I would prefer going by train unless it was unusably slow (think Amtrak to Montreal, not Amtrak to Buffalo); my mother would be more inclined to drive, especially if the trains were mediocre like in Italy; my father would drive unless the train was vastly faster.

  10. Matt T March 24, 2011 at 2:13 pm #

    A lexeme is the idea behind a meaning of a word, and as languages evolve words can stay the same as the lexeme drifts behind it, especially for spectrum words. I once read (hazily remembered, so this might be wrong) that in Welsh grey and brown have the same name, and in Russian that light blue and dark blue have different names so there isn’t the same idea as in the English blue. So I would say the term you might be looking for instead of a category-error is a vexed-lex, but with the standard caveat – I just made that up.

  11. Danny March 24, 2011 at 3:48 pm #

    I have a high enough income to be able to afford two newish vehicles in my home…and I still feel like owning two vehicles is a financial burden. When I tell people I know that I would love to get rid of one of my cars and take a bus instead, I get one of two responses: 1) “Why? You can afford an extremely nice car!”, and 2) “I know! I hate having a car!”
    Of course if you haven’t guessed yet, those who give me response #1 are people who might have seen Margaret Thatcher on TV quite a few times in their life, and those who give me response #2 are people who are closer to my own age.

  12. G-Man (Type-E) March 24, 2011 at 3:56 pm #

    My wife and I share a car. We have enough money for a second car, but spend it on maintenance and moorage for a sailboat instead. The quality of my life and leisure is vastly improved by this choice.
    On some days my wife needs the car to work and run errands. On those days I depend on transit to get home (We often carpool in the AM when she drops me off at work.) On days when she does not need the car I have a choice to take it, if there are particular errands I am running that would not be feasible to accomplish in a reasonable time via transit. I can also leave the car at home and take the bus if I want to save the fuel and get some exercise on the short walk to/from home. If I take the car, then she may decide to take the kids to the library on the bus, because she doesn’t have a car available at the time she wants to go. On a really pleasant day, any or all of us may bike to our destination. As a transit planner, I am very confused about how to categorize myself.
    Yesterday I talked with a disabled transit patron about ridership vs. coverage and ADA impacts. Our discussion came around to the fact that, at least in the US, the disabled have access to paratransit service, but they often choose, and prefer, to take fixed route services if they are frequent enough and routed to the right destination. That’s because they are cheaper and allow greater flexibility (trips don’t need to be scheduled in advance).
    I know you’ve addressed this, but just want to point out the similarity to the fixation on 1/4 mile walk. It seems to be the geographic equivalent of the category-error in language. Too often I find mapping done with a hard border (based on straight-line distance, not real walking distance) and quantifications that ignore anything outside that border. It is treated as an absolute instead of an elasticity gradient and I feel like our decisions suffer for that constriction. I don’t understand where the standard emerged from and why it holds such sway even as we collect data regularly that could be used to determine our own local walkshed elasticity factors. Perhaps there is a limit to the value that can be derived from more and more detail and nuance, but I feel like we lean too much toward simplification of our data. We rely on industry standards so much that we often lose an opportunity for local experience to guide us.
    Is some of this due to the need for understanding by the public and elected representatives? Are they so unfamiliar with or unable/unwilling to process such complexity as professionals might desire?

  13. Noel March 24, 2011 at 4:18 pm #

    Jarrett
    I think you may be talking about a “false dichotomy” or “false dilemma” fallacy. http://en.wikipedia.org/wiki/False_dilemma

  14. Nathan Williams March 24, 2011 at 7:18 pm #

    I have access to a car, and could drive it to work, although (as I experienced recently) 7x as expensive to park for a day as it is to take the subway (without even considering the economics of my unlimited-use monthly subway pass), and no faster. Am I a “choice” transit rider, merely because I have more than one option, or does the extreme economic disincentive change my category?

  15. J Prouse March 25, 2011 at 3:46 am #

    False dichotomy comes to my mind, too. Black and white thinking, etc..
    Category error has a different meaning in philosophy – generally to refer to the mistake of treating one type of thing as if it were another type of thing. This usually involves confusing levels of abstraction, or confusing abstractions with concrete, physical objects. One classic example is, “The Prime Minister is in London, and the Foreign Secretary is in Paris, and the Home Secretary is in Bristol, but where is the Government?” – the fallacy is that ‘government’ is a higher order of abstraction than the people that compose it ie a different category; it is not, strictly speaking, anywhere. Category error is prevalent everywhere in our language – even saying “Mr Smith is a brutish man” is somewhat of an error because you are referring to your experiences involving Mr. Smith, not to his actual physical existence as a human. Here’s a relevant article I found: http://plato.stanford.edu/entries/categories/
    Imposing hard categories on the world comes naturally, though, because that’s how our brains evolved to function.
    I suggest not using ‘category error’ since it already has a somewhat different meaning, but maybe you can coin your own, similar term? Perhaps “Categorization fallacy” – the error of assuming that things must belong to one category or the other, when in fact things may exhibit features (or deficits) of multiple categories – and there are many different ways of categorizing the same things… or something similar.
    Good luck with the book!

  16. Jase March 25, 2011 at 5:39 am #

    Some alternatives for your term ‘category error’:
    Wrong division
    Mis-typology
    Indistinction

  17. Dave March 25, 2011 at 6:56 am #

    This piece reminds me of the book “The Ethics of Metropolitan Growth” by Robert Kirkman. The title of that book threw me off as well, and I don’t think we as a population have spent enough time developing a narrative around this relative and non-specific way of thinking to immediately understand titles pertaining to this kind of discussion… this discussion being: instead of jumping to a conclusion and deciding that something IS a certain way, letting it be what it is and not immediately labeling it and categorizing it. I think the real lesson from the book was for people to analyze the unspoken assumptions and values they have, and figure out where they are on the spectrum.
    The other piece to the spectrum is perspective… where you exist on the spectrum significantly changes your view of the rest of the spectrum. So recognizing it as a spectrum and not boxes is important, but it’s also important to recognize the view from your own perspective, wherever it is, is always skewed differently than anyone else at a different place on the spectrum.
    The best line is: “These categories are imposed on reality, not derived from it.” – That’s the real piece that most people miss… so many people assume that the way they see the world is in some way fixed or precise, and don’t recognize their own role in creating their own perspective of the world. Once you recognize how you take part in building your own perspective, you can begin to recognize the way your view is skewed along this spectrum.

  18. Hermann March 25, 2011 at 9:38 am #

    I was reminded of the problems of histograms in statistics. On the one hand you need them to estimate an unknown continuous distribution. On the other hand the choice of position and number of bins is delicate and can lead to very different results.

  19. Sean March 25, 2011 at 10:17 am #

    I like the suggestions for “False Dichotomy” – it’s a term I recognize and can understand immediately.

  20. M1EK March 25, 2011 at 10:23 am #

    I actually occasionally use the term “transit-dependent by choice” to establish that the person in question could afford a car, but chose to pass on it because transit service was good enough to allow them to do so.
    Problem is that like many concepts here, this spectrum concept ends up just a rationalization tool allowing transit agencies to hear what they want to hear rather than what they need to hear. In the USA in particular, if you want to grow the market for transit, you are competing against the automobile, period, and that competition had better be won on metrics other than simple cost, because the variable costs of driving are very low compared to the fixed cost of owning a car.

  21. Daniel March 25, 2011 at 2:28 pm #

    Isn’t all language categorization? We attribute to words “definitions,” which are widely accepted boundaries (like boxes) to the meaning the term can convey. Putting something into words always requires some simplification of reality, but without that process, there would be no way to transfer meaning from one person to another or logically process anything within our own minds.
    So I wouldn’t say that the error lies so much with forming the categories themselves, but with applying them to situations where more nuance would be appropriate. Likewise, getting too lost in complexity will preclude us from finding real patterns in reality that can help improve decisions.
    I would be a lousy carpenter if I measured every cut of wood to the millionth of an inch. And I’d be a lousy engineer if I rounded everything up to an inch. Different tools for different purposes.

  22. Michael Druker March 25, 2011 at 2:48 pm #

    I’m really not sure why you chose the choice vs. captive example. I agree that it’s a really fuzzy distinction, but I view it as more of a perspective in transit planning than a firm categorization of any actual individual (prospective) riders. So much of what I see as transit planning is based on engineering approaches that seem to assume fixed demands, perfect information, and that frequency is mostly a way of dealing with demand rather than a constraint on utility. Those things are not true of even the most transit-dependent, but they’re more true of them.
    Which is where the “choice rider” term comes in. If it helps, one can certainly find people who have more choice in their use of transit to verify that the term gets at some real differences in transit use. But the target isn’t the riders so much as the planners. It’s a cognitive aid to planning the kind of service that is competitive and attractive, not to any one category of people but to most everyone.

  23. Jarrett at HumanTransit.org March 25, 2011 at 3:01 pm #

    Michael
    I think that is planners want to talk about a spectrum of transit dependence, that's accurate and not a problem.  But I've learned that I have to pepper my speech with constant reminders that I'm referring to a spectrum rather than to boxes, because adjectives sound like boxes, especially if they connect with other emotional needs to categorise that people may have.  If you see yourself as a choice rider, it's comforting in a way to feel conceptually 'walled off' from transit dependent riders, rather than at different positions on a spectrum.

  24. Morgan Wick March 28, 2011 at 5:19 am #

    You asked for a philosopher to chime in with a better name, but as several people above have shown, and as I argued in a paper I wrote earlier this week for school, I think philosophers themselves tend to be prone to this, especially if they fell in love too much with Plato’s “forms” where a chair somehow “participates” in some fixed, metaphysical “form of chairness”.
    Unfortunately the way to address this tends to be to claim that words are completely arbitrary and have no reflection in the real world (someone mentioned Derrida before, and from what I’ve heard he’d fall under this or something like it). I like to think that black exists and can apply to shades that aren’t (0,0,0), white exists and can apply to shades that aren’t (255,255,255), but the shades of gray can exist as well. To call some people “poor” must surely be accurate, if degrading, but for other people it could be argued either way, or you could create any number of new distinctions and classes.

  25. Art Busman March 29, 2011 at 2:49 pm #

    I think there’s multiple errors within the “category” error including errors of parameter, exclusivity, and over-generalization. Dividing riders into choice and non-choice makes them mutually exclusive, whereas, choices overlap. One can choose to walk or bicycle 10 miles and not be transit-dependent. A color can be purplish blue. Then there’s the whole one with the universe thing and concept of the self. It’s only a convenience that we think of our ‘selves’ in order to protect ourselves and feed ourselves, however, we can also think of ourselves as integral parts of a greater whole. The whole pro versus anti-transit thing can be misguided. What about personal rapid systems? What about time-shared personal jets? In nature, there really isn’t even such things as families, species, classes, it’s all just made up for convenience. Where do primates end and humans begin? What came first the chicken or the egg? It points out how deep philosophical concepts can skew our attitudes toward transit, how lobbyists can skew language and preconceptions about reality itself to sway the public.

  26. Art Busman March 29, 2011 at 2:58 pm #

    One of the best examples is the concept of race which is not scientific at all. What do you call a black Arab or a white Arab? What do you call a white man from South Africa? If he moves to America, is he African-American? There is more genetic diversity in a family of 50 chimps than the whole human race.
    In America, one assumes that a private car provides freedom, privacy, superlatives of luxury life, yet this is redefined. Suburban private life is viewed by many Gen Y’ers as stifling, imprisoning, isolating, alienating, even anti-social and disabling. Sharing transportation can be viewed as being in tune, attached, collaborative, networking. It’s all historical with early cities being over-crowded, unsanitary, prone to spread diseases. Thinking in categories, we also have to think about human geographic zoning and compartmentalization. I think we’re moving towards more co-mingling, co-habitating, co-ops, “co-mobility”.

  27. Kenny Easwaran April 8, 2011 at 3:01 pm #

    This issue is discussed by philosophers under the general heading of “vagueness” or the “sorites paradox” – if you have no grains of sand then you have no heap; if n grains of sand is too small to count as a heap, then adding 1 grain certainly can’t make it a heap; therefore, no matter how many grains you have it can’t be a heap!
    Timothy Williamson’s “epistemicist” view (which is growing in popularity, but I think most philosophers still reject it) is that terms like “heap” (or “blue” or “poor” or whatever) always do have a definite extension, so that we’re right about the premises of the argument – one of us is right and the other is wrong. However, he goes on to say that with these terms, it’s impossible for us to know exactly where the cutoff is (so it’s impossible for us to know the precise meaning of the words), which of course is why we shouldn’t really have these arguments, even though one of us is right and the other wrong.
    Linguists also point out that the terms that cause these problems are often what they call “gradable adjectives” (in that they have comparative and superlative forms), and that the semantics for these terms varies depending on context. That is, in each situation, the word “tall” applies to a definite class of people, given by a height cutoff that is (somehow) supplied by the context of our current discussion (say, whether we were just talking about children, or basketball players, or Americans, or Dutch people, or skyscrapers, or whatever). Given the context-sensitivity, we can’t take conclusions we’ve drawn about the term in one context and apply them in another context, unless we’re sure how the standards have shifted.
    As someone else pointed out earlier, the term “category error” is already established in philosophy (other examples would be when I think of the word “blue” as referring to a particular blue object, rather than the color it shares with other objects), so “false dichotomy” seems like one relevant term, or perhaps something about not noticing the phenomenon of vagueness.