Guest Post by Oded Goldreich

[Oded Goldreich has written an essay on the relative influence of “intellectual” versus “operational” goals in motivating our work, and on how this balance has changed in the past thirty years. In this guest post he tells us a bit about his methodology, his conclusions, and the changes he would like to see. — L.T.]

The purpose of this post is to call your attention to my essay “On the status of intellectual values in TOC”, which is available from the web-page

Before providing a brief outline of this essay, let me clarify a few issues.
First, the term “values” is adopted from Sociology, where it is defined as the set of beliefs of a society (regarding what is correct, good, and/or desirable). By “intellectual values” I mean a specific type of values; that is, those that advocate curiosity, study, and understanding. In particular, I believe that the TOC community holds (and should hold) both intellectual values and instrumental values. The issue at hand is the balance between them. Second, I am talking about intellectual values, not about intellectual activities; that is, I’m talking about what exists in the background. Third, I am talking about the TOC community as a social group, not about individuals who are members in that group; that is, I’m talking about the sociology of TOC. And last, my intention is to call for corrective actions, not to complain on the current state of affairs.

In the essay I study the status of intellectual values in the TOC community during the last three decades. Specifically, analyzing the motivational parts of papers that appeared in several STOC proceedings, I found evidence to my feeling that the importance attributed to intellectual values has declined in the last decade (or so). The said evidence is conditioned on a number of assumptions, which are spelled out in the essay.

I then discuss three theories that may be used to explain the decline of intellectual values in TOC (or rather three phenomena that may cause this decline). This discussion may be of interest also to readers that are unconvinced by my thesis and/or my empirical study, because it indicates potential dangers that loom over TOC.

The most intrinsically oriented theory asserts that intellectual values play a bigger role in the early stages of the development of a field, a stage which is marked by many works of explorative nature. In the essay I explain why I do not believe in this theory in general, and point out that it fails to explain the specific data that I gathered. Instead, I suggest to seek the causes elsewhere, specifically, in sociology. Two sociological theories seem most applicable here:
The first refers to the dynamics of the field (i.e., TOC) itself, while the second refers to the effect of society at large.

The first sociological theory asserts that as a field become more successful
(or, actually, is considered so from the outside), the competition within the field intensifies, and this creates pressures towards “objective” measures of accomplishment that can be reviewed from the outside. Such measures are typically oblivious of intellectual contents. Thus, under the reign of (externally monitored) competition, intellectual values decline.

The second sociological theory refers to the effect of the Zeitgeist on any activity that takes place in society (including scientific research). Specifically, the claim here is that intellectual values are in decline in the Western society for more than one hundred years, and that the decline has become more and more drastic with time.

Although my real objective is to advocate a restoration of the intellectual values in TOC, I believe it is helpful to study the past as well as the forces that might have affected it and may affect the future. In particular, the claim that things were different in the past provides some evidence that they may be reversed in the future. I admit that opposing the social forces that cause the decline of intellectual values is far from being easy. But I think that such an opposition is possible, especially since the TOC community is relatively small (which facilitates the creation of solidarity and the effecting of change). If the TOC community is determined to change its culture, then no outsider can prevent this. The outsiders will have to adapt to what the TOC community values; they have no choice (i.e., there is no alternative TOC community). It is only up to us!

In order to avoid claims of being too lofty, I provide a few concrete suggestions for the defense and promotion of intellectual values in TOC. These suggestions refer to actions that individuals can take, but they will be effective only if these individual actions will become sufficiently common.

  • Let the intellectual values guide you in your own research and in your interaction with other researchers.
  • When presenting a scientific work, provide an explicit account of the (current) motivation for this work.
  • When serving on either a PC or a hiring/promotion committee, try to steer the committee towards taking decisions on the basis of a real understanding of the contents of the work being considered rather than on the basis of some superficial “objective-looking” measures.
  • Object to the dominance of vulgar competition wherever it emerges.

Indeed, individual actions may be much more effective if they are socially coordinated. Thus, it may be useful to make these actions a topic of social discussion, to form groups that are committed to promote them, to create forums that promote them, etc.

A final note: Due mainly to technical reasons, I expect not to be able to participate in possible discussion that may evolve in this blog. Thus, if you wish to communicate with me regarding these issues, please write me directly (via email). (I will not object if you later post our correspondance.)

Oded Goldreich

69 thoughts on “Guest Post by Oded Goldreich

  1. I have just completed a sociological study of Oded’s essays over the past 30 years, classifying them according to whether they are explicitly or implicitly “crotchety.” I cannot assign a cause, but it is quite clear that while the implicit crotchety factor has remained steady, explicit crotchetyness has increased in lockstep with the decline of Western intellectual values.

  2. It is easy to mock Oded’s essay and there are many points in it that I disagree with strongly but I do agree that something significant has changed.

    I attended STOC 1983 as a second year grad student and, not knowing many people, I sat through about 3/4 of the talks. (I could do this because it was single-tracked but that isn’t my point.) Though the technology for talk delivery was not up to that of the Keynote, beamer, and Powerpoint talks today, I could follow a large portion of the content of the talks.because the authors made a concerted effort to make the motivations, results, and perspective for their papers accessible to a general theory audience. Talks were not just directed at the experts in the room. The same can be said of most of the papers.

    As a field, the diversity of topics in TCS has grown dramatically. This has been great for the vitality and impact of the field but authors seem to be losing sight of the need to communicate to their fellow researchers in the field as a whole and not just to a narrow audience of experts. (I have sometimes failed in this myself.) In communicating more broadly, authors need to express not only how their results are worthy in the context of other work in their specialty but also the motivations and directions that underlie that study as a whole. These latter motivations may be well understood by the experts but they need to be communicated to the field as a whole. (And indeed we also need to better train our students to answer where their work figures in the entire picture of the field)

    During my most recent STOC PC experience I was struck by how many more submissions than I had seen for previous committees failed to pass the test of being written for more than just the experts. For the FOCS 2009 submissions, the CFP required submission of separate 2-page summaries for a broader audience that I would previously have expected as a matter of course. There was some outcry and I am not sure that it could have been very useful since the requirement was not repeated.

    Nonetheless, the change I would like to see in the community is a re-dedication to communicating to the field as a whole, to making things accessible not to merely to the experts, to doing more than paying lip service to motivations and hiding behind prior related work as justification.

  3. It is difficult to know what to make of Oded’s post, because it is too vague, and I do not see any concrete criticisms in it. What exactly constitutes a “decline in intellectual values”?

    What I hope Oded means is that our focus has become on “the proof techniques” rather than the scientific and philosophical significance of our results (and yes, the techniques used to prove them at times). Even if that’s not what he means, I will go off on a rant about it anyways. A typical talk at STOC/FOCS, as well as the associated paper, can often be caricatured as “Hey everyone!! Look at this really hard math I just used / developed. Isn’t it awesome?!” Nowhere does the author pause to explain what their result, or their techniques, tells us about computation at a conceptual level. Essentially, techniques have come to be treated as an end unto themselves, rather than a means for understanding computation. As a result, a majority of our young researchers are “technicians” who are incapable of engaging most other researchers outside theoretical computer science in a conversation. This seems problematic.

  4. David, you are misreading. Research in algorithms, as in any other area, can have motivations that Oded classifies as “intellectual” (wanting to understand something new) or as “instrumental” (wanting quantitative improvements and/or difficult proofs). Considering Oded’s interests, I wouldn’t be surprised if he thinks that complexity theory is the worst offender in valuing the latter type of motivations too much over the former, and in being a symptom of the general decline of western civilization.

  5. Well, if I’m misreading, it’s because he includes algorithms research on a footnote on page 5 as an explicit example of something that he would consider as “instrumental”. Of course, the context of that sentence is that of algorithmic research that would lead to a practical improvement. I guess it’s ok to be curious about algorithms as long as you make sure only to study useless ones.

  6. The difference between “intellectual” and “instrumental” can be quite subtle. The essay is missing an Appendix listing papers and, for each paper, the sentence or expressions that determined his classification. Then we would have a chance to understand what it is that he is measuring.

    Without that, this is much too vague to have much value.

  7. Claire made the first comment that I would have made. At a minimum, I would expect a list of how papers were categorized, so others could attempt to reproduce the experiment faithfully. A pointer to specific word phrasings that led to that categorization would be standard. I’d go further and say that Oded should have had at least two others do the experiment with him, independently, so the 3 categorizations could be cross-checked against each other.

    I admit my personal opinion is that even then you’d want to check such human-based categorization with an automatic system based on word-usage so as to make everything repeatable.

    But Oded didn’t mean for this to be a scientific paper, he meant it to be an opinion piece, backed by “back-of-the-envelope” calculations. So perhaps this type of criticism isn’t useful, and instead we need more reactions like David Eppstein’s, which seems to be saying that Oded’s opinions just start from a faulty place. Or we need just more other opinions, whether agreeing, contrasting, or complementary, to get a sense of what others think about this purported problem.

  8. Goldreich’s essay is certainly a bold and impressive piece of work. I identifies with most ideals and goals in it.

    It is obvious that there is a decline in the value attributed to any intellectual activity in the West (and in most non-Western societies, there is no decline, because there was no value to it from the start).

    It is also obvious that this culture of “instrumentalism” or in fact “materialsim” (which I interpret as attributing values to “applications”) is in clash with any intellectual activity (in both sense: i.e., intellectual value and intellectual activity (here I go even further than Goldreich)); and it is also obvious that this culture endangers ToC, and in fact any scientific activity aimed at building a theory–hence, a framework of understanding.

    Nevertheless, one ironic thing about the essay is the use Marxist citations. First, Marxism is mostly an anti-intellectual ideology. Second, the actual citation contradicts in itself the goals set out by Goldreich: essentially, Marx states there that the aim of philosophy is not to understand reality, rather to change it! But isn’t this precisely what a ToC instrumentalist would answer to Goldreich’s saying that the purpoes of ToC is to understand!

  9. @David,

    Improving epsilon in an algorithm that no one uses in practice? I would like to know how many of the algorithms papers published in STOC/FOCS/SODA have practical use if that is what you claim. I guess that Oded would classify an algorithm that would create a significant increase in our understanding of the problem and make a major difference in practice as intellectual rather than instrumental. Having intellectual value does not imply being useless.

  10. @ToC researcher,

    “and in most non-Western societies, there is no decline, because there was no value to it from the start”

    Ignorance on your part about existence of something doesn’t mean it doesn’t exist.

  11. I guess it’s ok to be curious about algorithms as long as you make sure only to study useless ones.

    There are two problems here. One, is that you have jumped from the statement that research should not *necessarily* be directed toward “usefulness”, but can justify itself entirely by “intellectual” means, to the conclusion that it *should* not be “useful”. Your jump has no basis.

    The second, is even more important: what do you mean by “useful”? You probably can justify using the word “industrial” or “marketable”. But not “useful”. E.g., I don’t think that most current internet applications are useful. Social networks for instance are not useful. Or at least not only useful, and their absence is as “useful” as their existence. So why should we engage in research directed to social networks?

  12. As Claire said, Oded’s expression of how he classified “intellectual values” in paper introductions is obviously too vague to be of objective value. However, Oded would not have needed to list things on a per paper basis for this aspect to be largely checkable. If he had listed more specific language constructs associated with his classification and their frequency in the papers it would already have been checkable.

    For example: Appeal to authority rather than direct argument for justifying importance of a problem:
    “We solve/improve A which (is important because it) has received much study in papers X,Y,Z.”
    I’d bet that frequency of this has grown which I would agree supports his point that this is a loss to the field.

    On the other hand, if he specifically listed counts of “economic scenario as justification” as representing an instrumental value then it would mark large portion of AGT and mechanism design as instrumental. It would be no surprise that instrumental values classified this way would have grown and the “decline of intellectual values” in the 2000’s would coincide with the rise in popularity of AGT, which is surely not a loss to the field.

  13. I don’t know if what a paper says their motivation to be can be used to ascertain the true motivation behind the paper. I recently worked on an algorithmic problem out of curiousity. I thought a particular objective should be a fundamental objective and I wanted to know if current techniques could be used to approximate it, or to get anything better than random. After I developed new techniques for it, my paper was rejected four times by the mainstream conference organizations. Each time I had submitted a new, augmented version of the paper. Then, after essentially giving up, I accidentally stumbled on a very fundamental application. So I re-wrote the paper around this application and the paper got in to a mainstream conference.

    If you read the paper, you would think that it has “instrumental” motivation, because it has an engineering application, which I went out of my way to use as the motivation. But in fact the whole project was actually an “intellectual” motivation. That is no longer mentioned anywhere in the paper, because it did not help the paper be successful in any way.

  14. David, I don’t think that “it’s ok to be curious about algorithms as long as you make sure only to study useless ones” is the distinction that Oded is trying to make.

    If I were to describe to someone who doesn’t know about it why the work on the smoothed analysis of the simplex is important, I would say that it explains why the simplex works well in practice, and gives precise evidence to the intuition that bad instances for the simplex are fragile. These are “intellectual values.” If this hypothetical interlocutor asked me “yes, but do they give a faster LP algorithm?”, or “yes, but are their proofs difficult?”, then these would be “instrumental” values.

    Most of the work on the analysis of (existing) algorithm aims to “explain” rather than “do,” so its motivation is in the “intellectual” category.

    This being said, I don’t find this distinction very useful, because I think that, ultimately, the way I know that I have successfully “explained” something is that I am able to “do” something new with this explanation. (I made a related point in an old post) Maybe Western civilization will have to bounce back without my help.

    To me, the most interesting part of the essay is the conclusions, with the concern over the use of ‘quantitative’ measures to evaluate research, especially in light of what is going on in the UK. I don’t perceive this to be a problem in the (American) theory community, but I don’t know how things were before 1998 and maybe there has indeed been a change.

  15. I think most comments made here were addressed in the essay, but let me re-iterate a few points.

    1. The diff between “intellectual” to “instrumental” is explained at the beginning of the essay, and it makes no reference to any specific area of TOC. (Thus, I strongly object David’s reading! (*)) It is also clarified that the two “faces of science” are closely related, and yet different, and that I see no fault in the instrumental face. I’m concern of the possible decline of importance given to the intellectual face and values.

    2. I am well aware of the gap between the historical motivation and the motivation presented (which at some point I call “ideal”). I have no issue with researchers that decide to present an instrumental motivation in order to increase the chances of their paper to be accepted; I have an issue with the culture that promotes this. That is, I ask why does such a presentation increase the chances of the paper to be accepted?

    3. Just for the record: The interpretation of Marx’s quote and the labeling of Marxism as anti-intellectual is in *my opinion* incorrect.

    4. As I admitted, my study suffers from great drawbacks, which are explicitly stated in my essay. (But I find it unfair to call it a “back of envelop calculation”, since I spent about three weeks on it…) One of the problems, admitted in the essay, is that I used a classifier while being unable to extract the classifying rule. This, in my opinion, is not the worse problem, and TOCers should be well-aware of this situation (which is mentioned a lot in COLT research). I did compile an annotate list of all papers I read, but decided not to publish it because *I believe* that such a publication will make more harm than good. (See, e.g., David’s abuse of the example I gave in a footnote (*)).


    *) See Footnote (10), which lists four examples, including the one abused by David. My own intention was to say that *if* that is the statement in the paper, then I take this statement as of of an “instrumental” motivation. And note, again, that — in my opinion — there is nothing wrong with such a motivation!

  16. Let me add another couple of comments.

    5. The issue at hand is whether the current status
    of intellectual values in TOC is satisfactory.
    Paul and others (incl myself) have argued it is not.
    I’m most alarmed by the personal experience of “Occupy Theory”;
    if researchers feel that an “instrumental motivation” increases
    the (social) acceptability of their work, then that’s a reason to worry.
    I wish to stress that I do not desire a situation where claiming
    an “intellectual motivation” increases the acceptablility,
    I just don’t want it to be decreased.

    6. I think it is unhelpful to label people, and to discard
    their claims based on such a labeling. In my essay, I cited
    both Durkheim and Weber for things I learned from them,
    although they held quite opposite positions regarding most issues.
    Btw, regrading label, a friend of mine has just commented that in
    the last decade much of my own research was algorithmic in nature,
    although some “core algorithms people” may not think so…


  17. All this complexity vs. algorithms debate in the CS theory community is a little bit beside the point; no one in the CS theory community does algorithms which may be practical any more, so any motivation behind any CS theory algorithms research is really all “intellectual”.

  18. I can’t help but wonder if it’s really the k-out-of-n conference acceptance model that is promoting such angst. It’s all very well to spend time understanding the true (or presented) motivation for an intellectual piece of work and marvel at the deep thoughts being uncovered. But a conference program committee has much more prosaic problems: how to fit 100 marvelous submissions into 50 slots. At that point, all the subjective ruminations in the world will not help: you have to make hard decisions, and it is an attempt (maybe misguided) at consistency that creates a preference for “objective” criteria like “what’s the application” and so on.

    In other words, to the extent that conference capacity constraints distort the value system of the community we have a problem (which is to say that we do have a problem), but the solution is (as people are observing in endless different ways) to provide many different forums for dissemination and evaluation of research.

    p.s A.Anon – at the risk of troll-feeding, I wonder if you’ve ever used a Bloom filter, cuckoo hashing, LSH, near-neighbor algorithms, distributed caching, data stream computations, or any kind of large-data analysis dimensionality reduction tool.

  19. And let me raise a question, and call it comment (7).

    Somebody has invested a month of his life
    conducting a study and writing a corresponding essay.
    (He says he spend 3 weeks doing exhaustive study,
    and you may add a week of fun in writing.)
    The essay has some faults; its style is clumsy,
    it fails to articulate some points clearly enough,
    it uses too many assumptions, etc.
    Why try to mock this person?

    Oh, I forgot that you think that he was trying to promote himself.
    Do you have any proof for that? Is there not evidence to the opposite?
    And if he was trying to promote himself and the idea is bad,
    will such an attempt not harm him instead?
    But if the idea is good and he also benefits,
    is that so bad? Can you not consider it a fair fee?

    Let me reassure you that I wasn’t really insulted,
    I’m just wondering about this behavior.

    Or maybe you feel insulted by my own essay and feel that I tried to insult you?
    In such a case, let me restate loud and clear that I did not mean to
    insult any person; my entire point was the social dynamics of TOC at large.


  20. Let me take one example: an asymptotic approximation scheme for strip packing.

    – If the introduction stated: “The previous best asymptotic approximation was a factor of 3/2; here, we improve that factor to 1+epsilon” – then it’s classified as instrumental, I think.

    – If the introduction stated: “We give a 1+epsilon asymptotic approximation, thus establishing that there cannot exist any hardness of approximation result” – then it’s classified as intellectual, I think.

    Yet it’s the same result…

    That’s why the difference is too subtle for me, and I would need examples to understand it better.

    I’m also a little bit taken aback by seeing “improved performance” classified as “instrumental”, because, P versus NP, isn’t that all about performance? Again, I would need examples to understand what is meant by “improved performance”.

  21. Let me briefly address Claire’s comment, without referring to the specific example.
    Yes, the very same result may be classified by my analysis as either “intellectual” or “instrumental” or as a “mixed”, depending on the motivation that the paper offers! Note that, indeed, these adjectives refer to the motivation given to the reader. The point of my essay is to ask of the change in this aspect over time, firstly, whether it exists at all, and secondly what are its causes. My main thesis is that the current *culture* of TOC discourages “intellectual motivation”, and that this in turn is caused by two social phenomena (not to be repeated here…).

    The question of whether a change in this culture has *already* cause a change in the profile of the research was not addressed in the essay! I assume we all agree that it is bound to have such an effect on the long term. But even without such a consequence, I am very concern of phenomena as reported by “Occupy Theory”; I have heard and seen much of that in the past (*), but it was handy to have this supportive evidence (although it was intended to refute my theses).


    *) A good question is how frequent are works that were truly motivated by intellectual consideration and/or with results mainly of intellectual appeal being presented by instrumental motivation, vs the frequency of the opposite.
    My clear impression from my study is that there are much more of the former type nowadays!

  22. Let me now address Suresh’s comment.

    My point is that intellectual values should be acted upon
    in real-life situations, not merely used as lips service.
    That is, experts should also evaluate the intellectual contribution
    of research, not merely its instrumental contributions.
    In both cases, this calls for understanding the work,
    rather than trying to rely on superficial objective-looking
    measures of unclear meaning (such as paper counts, citations counts, etc).
    This is doable in the real world, not in an ideal one.

    The “real world” (of TOC venues and procedures) is not a given thing,
    it is what has evolved in history by the agents of the TOC community,
    and it can be changed in a similar manner (i.e., by our behavior,
    especially when “socially coordinated”, which admitingly has
    also to overcome external social forces discussed in my essay).

    This being said, let me confess that I’m far from enthusiastic about
    the current state of our conferences, and especially of STOC/FOCS.
    As I see it, the original and well-motivated function of these venues
    is to serve as channels for dissemination and exchange of ideas and knowledge.
    But, as Paul noted, this function is largely lost in current STOC/FOCS,
    which currently serve merely as “weight-lifting competitions”.
    That is, authors do not really make an effort to communicate their ideas
    to the TOC community at large, the PCs don’t really insist on this,
    and the audience no longer expects it. Everybody seem merely concerned
    with which papers deserve the “award” of being published in STOC/FOCS,
    and which papers got that award.
    This being the situation, I lost all interest in submitting papers
    and/or attending these venues, and I think TOC will be better of
    if they are abolished (if not restored to their original function).


  23. One more comment, not on the specifics of the essay but on the general complaint about decline of intellectual values.

    I agree with Paul Beame that much has changed (since 1986, the first FOCS conference I attended.)

    Regardless of how introductions are written, I find that the field is much deeper. What do I mean but “much deeper”? Take a random mathematician, give him some background information, and have him sit at a talk at STOC or FOCS: twenty years ago, my impression is that he would have been able to follow the talk, read and understand the paper, after perhaps an hour’s worth of introductory material. That is no longer true. Now, to understand what is going on (particularly around the UGC), my impression is that he would need a semester-long course: 26 lectures instead of one! Why? Because the results use a much more versatile palette of more sophisticated technical tools, bringing together diverse fields. I view that as something to be admired rather than deplored. yes, a lot of what recent PhDs are doing is above my head: rather than complain about and communication, I admire their knowledge. (I guess I can still complain about communication, but that’s a secondary concern.)

    My assessment is that the field has matured. It is deeper. Even if we restrict attention to pure Theory, that is now a full-fledged field which takes years to master. And, in my view, that is good. It is a resounding success. We are building an impressive theory of computation.

  24. Personally, I’m not that keen on the adjective “deep” nor do I take the amount of background material required as a measure of great importance. What I care about most is interest and importance. Is research nowadays more or less interesting and/or important than a decade or two or three ago is a question I did not entertain. For sure, I see recent PhD students that understand things that I don’t, and I deeply respect them (well, here I just used “deep”, but applied differently, I think…).

    All this has nothing to do with my concerns and with my essay, which may mean that my essay does not address the most important or interesting social question regarding TOC…. Still, my essay addresses a certain question, which I find important and interesting, as well as some questions derived from them. Forget my essay, here are the questions.

    1. Is there a decline in the status of intellectual values (not of intellectual activities!) in TOC?

    2. Is this good or bad?

    3. Assuming we agree it is bad, can we do anything about it?

    4. Assuming we can try, what should we try?


    p.s.: And do note that the *responsibility* for the social aspects of TOC, as reflected in the above questions, rests with the more senior researchers. This is not said to exclude others, but rather to assert that if blame is to be distributed (which is not my intention at all), then it will go to the seniors. (I’m saying this because I have a vague impression that Claire felt an urge to defend the less seniors, which I did not mean to attack at all…)

  25. p.s.: Or consider a set of questions that are closely related to the opposite of the four questions I just listed:

    1. Is there an increase in what I called “vulgar competition” in TOC?
    By this I mean an *excessive* preoccupation with competitions of various types (e.g., the current way that STOC/FOCS opertate and are viewed)?

    2. Is this good or bad?

    3. Assuming we agree it is bad, can we do anything about it?

    4. Assuming we can try, what should we try?


  26. @Claire,
    As I said before, I completely identify with Goldreich’s agenda.
    I didn’t fully understand what is the relevance of your comment about the “deep” nature of ToC. I actually understood Goldreich’s thesis as attacking in some sense the “superficiality” in ToC. A lot of “instrumental” results, or “instrumental motivated” results in ToC are actually not “deep” in the mathematical sense at all.

    Weightlifting per se is certainly not something to be admired. But in my opinion, seeking deep mathematical understanding of some (even technical) questions, is, and should be, considered as an admirable”intellectual” pursuit

  27. ToC researcher, it seems that Oded is not attacking the “superficiality” in ToC, but the superficiality of the values expressed in the introductions, independently of the results themselves. (That’s why, as Oded said, my last comment was shifting the discussion.)

    He is criticizing form, not content. The same content will be classified differently, depending on how it’s presented in the introduction. He is only assessing the wording, the particular turns of phrase, the perspective chosen to present things in the introduction. He does not take into consideration the result itself. (Oded, correct me if I am wrong.)

    He claims the form reflects something more meaningful that he calls the “intellectual values” of the field, and that his classification captures something of that. But I cannot be convinced of that without specifics. I need to see those wordings and those turns of phrase that triggered his classification.

    I wonder. If he applied his criteria to the FUN workshop, would those papers turn out to be 100% on the “intellectual values” side and 0% on the “instrumental” side?

  28. I asked myself whether the motivation presented to the reader is dominantly intellectual (i.e., seeking understanding or being curious) or dominantly instrumental (e.g., seeking to improve performance and/or address practical concerns).

    My biggest problem with Oded’s study is that it proposes a false dichotomy. The distinction between “dominantly intellectual” and “dominantly instrumental” reeks of the endless (and pointless) arguments about the relative merits of “pure” versus “applied” mathematics. Where a particular work lands in the spectrum is at least as much a function of the reader as the author. One reader’s “instrumental” is another reader’s “intellectual”, just as one reader’s “pure” is another reader’s “applied”. See footnote 24.

    I think this is the core of David’s objection: Oded explicitly labels a particular species of motivation for TOC research (“Is there a faster algorithm?”) as instrumental and therefore inferior, which David both values highly and regards as primarily intellectual. (I hope David will forgive me if I am inaccurately projecting my own values onto him.)

    The seriousness of this (unintended) insult is inflated by the contention that the shift from intellectual to instrumental values reflects a similar shift in Western society, again, with the clear implication that this shift is A Bad Thing. Hence David’s snark about “the decline of Western civilization”.

    But I find this contention strange; when was Western society ever based on intellectual values? The illustrative example of “people use Google without understanding how it works” is particularly bizarre. Did “traditional societies” really have more than a crude functional understanding of the tools they used? Really? Consider basic examples like fire, or agriculture, or language, or soap, or money, or steel, or electricity, or medicine, or television.

  29. @Claire,
    I think that you may not be reading correctly. The motivation and concern is certainly a shift in the content, not merely in the form. It might be an artifact of the methodology taken by Goldreich, that his essay attacks the “form” but not “content”, but this is not what is meant, or is supposed to mean. And this can only show the limitations of the techniques taken by Oded.

    As Oded writes, he is interested primarily in the content (=”profile of research”):
    “The question of whether a change in this culture has *already* cause a change in the profile of the research was not addressed in the essay! I assume we all agree that it is bound to have such an effect on the long term.”

    So, I don’t think we should pick on the methodology, and take extremely formally this “sociological text analysis” (here I may disagree with Oded). We should consider the intent of this essay, and not analyze it, since it is an opinion essay more than a research paper.
    And for this matter, I’m interested in your opinion, Claire, wouldn’t you say that what Oded tries to prove, i.e., the shift from predominant intellectual to mainly instrumental motivations, is occurring in ToC?

  30. Well, as I have indicated before, I think that in terms of structural content, ToC is doing great. I am proud to me a member of the TCS community, and feel lucky to be in a field where so much is happening.

    If I have any concerns, it is about the division between TCS and the rest of the world. When what we term “algorithm” has little to do with programs that people write, and when “efficient computation” has little to do with fast programs, I am afraid that we are depriving ourselves of the richness that can come from confrontation with reality, and that we may be limiting ourselves to a world that is of our own construction, with very few links to the outside. It’s not good for funding or for our image, but more importantly, in the long-term it might not be so good for the intellectual richness of the field. There is a gap that has become almost a chasm by now, as the field has evolved; so I think that trying to maintain and strengthen ties with computation in the real world is a priority.

    Unfortunately the real world is too messy for my personal taste; I get impatient with the kind of obstacles raised when designing programs and when trying them out on real problems. So I am left with mere good wishes for those researchers who get excited by implementations.

  31. @JeffE,

    I don’t agree with your statements.

    1. You claim the “spectrum” between intellectual and instrumental is subjective, or imaginary. First, I don’t see how you can justify this claim. The fact that a distinction is not completely crisp doesn’t mean it does not exist. If I set out to understand the origin of life, for instance, or the creation of the universe, versus, if I set out to build a better mechanism for site rankings, I think that the distinction between the intellectual and instrumental is rather clear.

    2. I think that the essay explicitly says that there is nothing “inferior” or “superior” here, but the concern is to maintain the balance, or to accept that either the intellectual or the instrumental motivations are legitimate. So the concern is to correct the current move into the predominant instrumental motivated research, not to claim it is inferior.

    3. I’m not an historian, but in some, even contemporary cultures, “intellectuals”, i.e., those that are involved in theoretical research directed at understanding fundamental phenomena, regardless of their applicability, are regarded pretty high, and their goals are regarded as noble. In some cultures they are not. We are certainly moving from one direction to the other direction (at least this is my impression). This what, I think, the essay means by “decline in intellectual values in the West”. Note also that the essay does not use the term “decline of the West” rather “decline of intellectual values”, so you cannot interpret this statement as a dichotomy between “inferior” and “superior” aspects of civilization.

  32. This is an excellent article in that rather than spout off things that are
    `obvious’ he actually gathers data. I find the data compelling.

    As for his suggestions, here are my comments:

    1) Let Intellectual values guide you. AGREE that we should be AWARE of
    how our research fits into the grand scope of things and we should know what the motivations are. MILD DISAGREE- there are problems you just happen to come across randomly and nothing wrong with pursuing them.

    2) We should MOTIVATE more. This is obvious but NEEDS TO BE RESTATED TIME AND TIME AGAIN!!!!! The FOCS 2-page-of-motivation
    rule seemed like a good idea (why did they abandon it?). Papers in all of the major conferences really need more motivation. Even if the motivation is
    that this MIGHT lead to a result that MIGHT lead to a result that MIGHT…
    its worth stating.

    3) PC committees should take intellectual values seriously. Probably agree
    but not quite sure how to go about this. I feel strongly that they should take
    MOTIVATION seriously. And also how well-written a paper is.
    I thought the INNOVATIONS conf was supposed to help this problem- did it?

    4) Object to Vulgar Competition (e.g., awards and paper counts). When someone comes up for Tenure of Full Prof it is typical to say
    `and she has 117 STOC papers!’ Should we instead say
    `She has 20 papers that have intellectual value’ ? Not sure.
    I DO think that LETTER WRITERS should address QUALITY of papers,
    and intellectual value.

    DISCLOSURE: I’m not quite sure what `intellectual value’ means
    and it is hard to define.

  33. I don’t understand the implication (from Oded and ToC researcher) that we currently need to “maintain the balance” from, say, 20+ years ago.

    Things are a bit different today. Computing is ubiquitous in ways it wasn’t even from the time I was in graduate school. Why wouldn’t that change both what we research, and our motivations? Perhaps we are more instrumental rather than intellectual right now, I don’t know. (I share Claire’s objections to how the problem has been framed and examined thus far.) I also don’t think I’d be worried if that was the case.

  34. @Michael,

    Well, for this I would answer with an ideological and an historical argument: first, any basic science needs an intellectual basis; that is, a science or a theory is something that is meant (a priori) to explain and enable the understanding of certain phenomena. If you change the balance (as you seem to implicitly admit), then you are not doing science, but you are left with an instrumental, or engineering-type, field. So “by definition” this is not the “purpose” of science.
    Of course, this is something you might object. But now comes the “historical”, argument: even if you think an instrumental-dominated field is the right way for ToC, then we should remember that some scientific disciplines that were completely instrumental didn’t survive, or have been degraded or needed a drastic change in order to recover from this state. An example that comes to my mind: ancient to mid-age astronomy (I’m not an historian, so I cannot come with a better example currently).

  35. What I feared initially (when I said that I may not be able to participate) is actually happening, but I am comforted by the “TOC researcher” who seems to say many of the things that I’d want to say.

    Regarding the questions of change, balance, and desireable, I wish to call attention again to the report of “Occupy Theory”, which I believe to be representative: A person reports that an introduction of one type was modified into an introduction of another type in order to increase the paper’s chances of being accepted in a conference. Should we not be shocked by this? (Regardless if the 1st was intellectual and the 2nd instrumental, or the other way around!)

    Regarding what should be done in committees, I of course do not say that they should only read the motivation. They should read the paper, understand the results, and evaluate them. You may say that this is what is being done anyhow (in PC and promotion committees), and I’d answer that I heard too many references to other issues (e.g., various counts mostly in promotions, bluntly superficial evaluations in PCs, etc).

    And PCs should insist that the paper provides a motivation, as part of an introduction (of say 2-3 pages) that should be addressed to a typical TOC member (rather than to a handful of experts!). If the authors cannot do that, then what is the point of having the paper in the conference? What will they actually present at the conference — a talk intended to a handful of experts? If so, why ask more than 30 other people to attend this talk???

  36. Being a very young researcher I don’t feel capable of having an opinion on Oded’s claims, but I’d like to try to explain how I understand Oded’s essay (Even though I do not agree nor disagree with it).

    I think that it is easier to understand the difference between “intellectual” and “instrumental” by looking at Biology rather than ToC. Consider two simplistic examples of biologists:
    a. The first biologist studies the human body because of mere curiosity – he has a passion for knowing how the human body works. Such a biologist would be considered “intellectual”.
    b. The second biologist studies the human body because he wishes to find a cure for cancer. Such a biologist would be considered “instrumental”.

    Is the “instrumental” biologist inferior? Obviously not! His research is likely to be more important than the “intellectual” biologist’s research, at least in the short term. As I understand Oded’s essay, this is also the way he views instrumental research.

    Oded’s point, if I understand correctly, is that a scientific field can not rely *only* on “instrumental” researches such as the research of the second biologist above, and also needs “intellectual” researches such as the one of the first biologist above.

    Oded claims that the “intellectual” type of motivations in ToC is in decline. More specifically, he claims that regardless of the motivations of individual researchers, the current culture of ToC discourages “intellectual” motivations, and researches that are motivated by such motivation.

    To support his claims, Oded argues that over the years, while the fraction of papers that are explicitly motivated by instrumental motivations has remained roughly the same, the fraction of papers that are explicitly motivated by an “intellectual” motivation has decreased sharply, and on the other hand, the fraction of papers that do not motivate their research at all has increased. Oded’s conclusion is that today’s researchers feel that it is not socially acceptable to use “intellectual” motivations for their papers, or at least, not as socially acceptable as it was in the 80s.
    The comment of “Occupy Theory” supports Oded’s claim, telling that he had to change the motivational discussion of his paper from an “intellectual” one to an “instrumental” one in order to get the paper accepted.

    Oded concludes by arguing that this situation should be remedied, and suggests few ways for doing it.

  37. ToC writes: If you change the balance, then you are not doing science, but you are left with an instrumental, or engineering-type, field.

    Actually, outside of physics and math, most sciences do not make a distinction between their engineering and theoretical branches. In medicine, biology, chemistry you will find both people doing high level theory and rather specific research (how to cure disease X or produce better plastic Y) even in the absence of an underlying theory.

    It is worth noting that even within physics and math this distinction is mostly a late XIX-early XX century notion. Up and until then, mathematicians such as Gauss, Poincare, the Bernoulli’s, Elie Cartan, Hilbert and Euler readily jumped in the waters of instrumental applied work seeking intellectual challenges and ideas.

  38. And PCs should insist that the paper provides a motivation, as part of an introduction (of say 2-3 pages) that should be addressed to a typical TOC member (rather than to a handful of experts!). If the authors cannot do that, then what is the point of having the paper in the conference? What will they actually present at the conference — a talk intended to a handful of experts? If so, why ask more than 30 other people to attend this talk???

    It should no surprise that I agree completely with Oded’s comment above.

    Claire: I agree that TCS now often uses sophisticated techniques that were not used back in 1983 but should that excuse authors from the above? Is “narrow-casting” of papers to the experts on the PC enough?

    Given the explanatory power one gets if the UGC is correct (which you cite as an example of a sophisticated topic), I don’t think there is a legitimate excuse for not doing more.

    (BTW: STOC 83 papers used a large amount of sophisticated math that I have not seen used recently – just take a look at the program.)

  39. Without entering a historical debate with Alex re other fields, let me stress that the way I refer to the distinction between “intellectual” to “instrumental” is conceptual, not institutional nor personal. Specifically, looking at any work, I can ask what values does it promote (or what face does it serve), and the answer may be A or B or both… This does not mean that individuals or groups are divided in this way (e.g., saying that some of my works are “intellectual” and some are “instrumental” is irrelevant). This conceptual division reflects things that are going on in reality, in the evaluation of individual works (either in formal or informal forums; e.g., in PCs and in the classroom), which in turn may affect future research.

    The issue is the importance assigned by the community to the two “faces” and value-clusters. I claim that too little importance is currently assigned to intellectual considerations and values, and that this is bad. This is not an “academic” discussion about semantics, but rather has very concrete consequences (as examplified by “Occupy Theory” story).


  40. 1. You claim the “spectrum” between intellectual and instrumental is subjective, or imaginary.

    Definitely not imaginary, but very subjective and very fuzzy.

    2. I think that the essay explicitly says that there is nothing “inferior” or “superior” here

    Then why describe the perceived shift with such a tone of concern, or tie it (strangely, again) to a perceived rise in “brute competition”, or explicitly follow the analysis with “a few concrete suggestions for the defense and promotion of intellectual values”? No, the implication that instrumental values are inferior is quite clear.

  41. Paul, that’s a different issue. The discussions goes into at least three different directions:
    1- Do introductions motivate results using a style that emphasizes “intellectual” or “instrumental” values more? Has that changed over time?
    2- Is the field – the actual research activity – becoming more superficial?
    3- Are introductions and talks geared to experts or to theoreticians in general? Should they target a broader audience?

    I have criticized the essay by arguing that point 1, which I thought was Oded’s main point, is supported by unverifiable data.
    I’ve given a quick personal opinion on point 2 (without engaging in discussion abut it), but only because one commenter asked me for my opinion about it.
    I think that point 3 is a diversion.

  42. I agree with Jeff and Claire, in that I thought Oded’s essay was about “intellectual” vs “instrumental” values, and the decline of “intellectual” values over time. I don’t think Oded has provided verifiable data. I also don’t think he’s demonstrated a sharp decline in intellectual values in the field in any case. Personally, I view the fact that people make an effort to tie their results to “practice” as a positive for the field; given the state of computing it would be hard to see how it would or should be otherwise. I don’t see that implying that “intellectualism” in theory is in decline; people who are not so interested in practice and are motivated purely, mostly, or at least significantly by the intellectual challenge still abound.

    In short, I’m not clear there’s data to support Oded’s claims on this point, and even if there was, I’m not convinced I should be concerned.

    Others, including Oded, have somehow tied this into other issues, such as the rise of competition in our conferences, whether research is becoming more superficial, and how people are tailoring their writing and talks to other experts rather than the wider community. While each of these is a potentially interesting issue and cause of concern, I don’t see how they connect to Oded’s main point except in extremely tangential ways. That these issues have arisen so readily in the discussion suggests to me that they are issues of concern that should be further discussed, although I think the discussions of these issues would benefit if they were not tied to the essay, as they seem essentially independent of the original point of that work.

  43. There are an infinite number of questions we can ask, therefore we need a set of criteria to identify which of those are worth publishing the answer for.

    One criterion is pure curiosity (what Oded misleadingly calls intellectual value) another is applicability, a third one is technical difficulty. It is important to note that these three usually overlap to various degrees.

    In terms of the Occupy Theory anecdote, I don’t see anything alarming, there, at least in principle. Say someone proves a silly theorem about a specific group with three generators that no one cares much about. Naturally the paper gets rejected. Then, surprisingly that someone finds that this magic property allows you to break DES and the paper is published right away. Where is the scandal there?

  44. I think differently than Michael, JeffE, and Alex Lopez-Ortiz:

    1. Oded has provided basic and appealing data to substantiate his main claim: a decline in intellectual values vs. instrumental values in ToC.

    2. The distinction between intellectual to instrumental is clear, though can be sometimes fuzzy, like any other non-mathematically defined concept, but not fuzzier.
    For JeffE: I think that the distinction, e.g., between a “good” paper and a “not-so-good” paper is much more fuzzy and subjective than the intellectual vs. instrumental distinction. So as long as the conference system is ruled by committees aimed to select the “best” papers out of their submissions, we cannot claim that crisper distinctions, like intellectual vs. instrumental, cannot be applied or detected in practice as well.

    3. For Michael: The decline of intellectual values is certainly something alarming, as a continued decline might lead to an engineering-type field, which may well mark the possible end of ToC (I argued for this above).

    4. For Claire: the current successful and satisfactory state of ToC, as you experience, may be in fact attributed to the intellectual-oriented work done in the last ~30 years in ToC. So the actual harmful effects of the decline will possibly be seen in the future. Not now.

    5. For Alex: Intellectual values are (or should) not be attributed to curiosity alone. I prefer (though this is my view) to call something “intellectual valued” if it aims at understanding a phenomena and not at devising instruments (or solutions to a specific problem, while lacking an explanatory nature).
    “Occupy theory” states explicitly that the intellectual motivations described in the paper “didn’t help in any way to the paper getting accepted”. And so he removed these motivations from the paper. This is a possible alarming indication of a decline in the intellectual value in ToC.

  45. 1. Oded has provided basic and appealing data to substantiate his main claim: a decline in intellectual values vs. instrumental values in ToC.

    He hasn’t even succeeded at communicating what exactly he means by the loaded term “intellectual value” much less substantiated his claims.

    2. The distinction between intellectual to instrumental is clear, though can be sometimes fuzzy, like any other non-mathematically defined concept, but not fuzzier.

    So the latest definition after several rounds of questioning is “explanatory curiosity”. Ok, let’s use this one for the rest of this post.

    The decline of intellectual values is certainly something alarming, as a continued decline might lead to an engineering-type field, which may well mark the possible end of ToC (I argued for this above).

    You claim as self-evident that the field would be better off it had more papers explaning phenomena for their own sake rather than solving actual problems. This is far from being a prima facie true statement.

    In my opinion the ideal is a combination of both. Oded implicitly claims, without even pausing to think about it, that the relative weight given to them in STOC a few decades ago was “The Right Level” (TM). From there he implies that any decline (real or imagined as his experiment is impossible to replicate) has to be bad.

    Maybe curiosity was overrated back then? or perhaps as the field changes solving problems becomes more important? or perhaps we are fortunate to work in a field where one can study important phenomena (like the power of two choices, or the limited space computations) while having applications? Or maybe people prefer to highlight applications as it is much easier to explain the value of an idea in those terms than in “big picture” this-might-be-useful-for-our-understanding terms?

  46. Oded,
    In the version of your paper on your website, I did not see a clearly marked section of references. I am particularly interested in references to analogous works of sociologists, and where I can find a sociological statement of some of the overall theses that you make, e.g., the general decline of intellectual values. I haven’t read the full thing in detail, but I find it hard to go from your numbers to your conclusions. What are reasonable error bars for significance for small amounts of data like you look at? I also think that it is strange that motivations such as practical utility and technical virtuosity are lumped into the same category. Are these words “operational” or “instrumental” commonly used to mean what you mean by them, and contrasted with “intellectual”? Or is it short-hand for “Things Oded doesn’t think are intellectual”?

  47. I reiterate that almost all the comments and/or questions raised are addressed in the essay. This includes loose definitions of the main terms, which may be found in the beginning of the introduction. These definitions are as loose as those you see in dictionaries and in non-mathematical discussions; so indeed one must adapt expectations when reading them.

    I think it is pointless for me to reproduce here text that can be found in the essay. I wonder whether some of the commenters have actually read the essay (or just the overview provided in my initial post). (I find it unfair to react based on the initial post, which indeed — due to its brevity — does not even provide a loose definition of key terms.)

  48. Oded, despite your effort, people are not simply going to trust the gross numbers of papers in each broad category that you list. I don’t see an alternative to including counts of the more specific categories of phrases and how you classify them. This would at least make your categorization others. If a paper has more than one category, how does that count in your scheme?

    The advantage of including this (which is the kind of thing that is typical in the sociology literature) is that people will be better able to make their own value judgements about the implications of the changes you cite.

    As for the other aspects and explanations you give, I recall plenty of “gamesmanship” and “competition” about STOC/FOCS submissions going back to my days as a postdoc at MIT in 1986-7. I do not believe the competitiveness has increased at all and the acceptance percentages are not so different from back then. (Indeed, shortly thereafter, the theory job market and competition for positions was much worse than it is now.)

  49. Let me clarify further what I meant to say in the last comment.
    I think it is reasonable and fair for people not to agree with
    my conceptual framework, definitions, assumptions, analysis,
    and conclusions. Likewise, it is reasonable and fair to think
    that what I provided is not good enough and/or is fundamentally
    flawed in that way or another. I really respect such critique,
    although I’m likely to disagree with it. What I find annoying
    and unfair are claims by which I did not provide one or more
    of the above ingredients.

  50. Dear Paul.

    I agree with your 1st point, and do not expect people to simply say “amen” to the analysis that I provided. I think it is totally reasonable and fair to object to any of the assumptions listed in Sec 2.3 of my essay.

    Regarding your specific questions in the 1st paragraph: (i) as stated in the essay, a paper presenting two motivations of different type (with none clearly dominating the other) was counted as “mixed” and appears so in the analysis; (ii) as stated in the paper, I cannot formalize my classification rule (and I believe that’s as infeasible as trying to formulate what makes a paper good); (iii) I do have a detailed analysis of approx 200 papers which records (for myself) why I chose to classify it as I did, but I chose *not* to publicize it (and I think I was right in doing so). As admitted in the essay, my own classification cannot be verified by others, but what others can do is carry out their own study, while using themselves as classifiers. I would welcome such a study, but do warn that I invested 3 full working weeks in my study….

    Let me just say explicitly the reason that I will not publicize my detailed comments regarding the papers. Firstly, I think this will be of *little* help wrt the reasonable concerns that refer to the validity of my assumptions. But more importantly, I fear that such a publication will evoke great hostility; just see what is going on in this blog and extrapolate.

    Regarding your last paragraph, I do agree that competition and various related aspects was very present also in the 1980’s. My claim is that it has increased significantly from that level (which I also found excessive…), and that the attention given to other aspects (i.e., the actual contents of the scientific work, and its motivation — especially when it is intellectual) has significantly decreased. So who is right? This is the question that puzzled me before I went into that study: That is, there was my feeling (supported by other researchers who felt the same) and there was your feeling (expressed by other researchers). The study I conducted has increased my confidence in the correctness of my feeling, but I cannot expect it to have a similar effect on everybody. I expect (and expected) that it may be instructive for some people (and at different levels). This is why I published it. In addition, I seized the opportunity to advocate the importance of intellectual values and to call for action aimed at promoting them; one can join this call regardless of one’s belief in the validity of my study…


  51. Of all the comments in this post, what my day is the consistency of use of CAPITAL LETTERS in GASARCH’S comment :).

    NOTE: This is *NOT* meant to mock or insult his style or comment. It is just something I noticed and smiled upon :=)

  52. Oded,

    I did not mean to say that you should publish the specific assignments on a paper-by-paper basis. That could indeed generate hostility. What I meant was one level above that but one level below your overall assessments: 34 papers focused their justification on practical near-term considerations. 26 papers referred to improvements on prior work as a main justification for importance.. 10 focussed on both. 21 described problems as stepping stones to answering bigger/more important questions.. You have aggregated these at one level too high for verification. In your essay you give examples of different motivations that fit your overall assessment but not at a level at which I as an outsider can be comfortable with.

  53. Alex:

    “Or maybe people prefer to highlight applications as it is much easier to explain the value of an idea in those terms than in “big picture” this-might-be-useful-for-our-understanding terms?”

    This is exactly what Oded is talking about:

    1. Oded claims that researchers feel that “in big picture this-might-be-useful-for-our-understanding” is no longer acceptable as a motivation for studying a question.

    2. Oded claims that this is a problem, since a scientific field must also study questions driven by this motivation (in *addition* to application-driven questions).

  54. Or, I’m actually saying something slightly different than Oded.

    I claim that explanatory curiosity is harder to justify per se even in the presence of unbiased refereeing. Whereas a claim of applicability is relatively objective, explanatory curiosity is subjective and often to identify the centrality of a question requires substantial familiarity with the topic. Needless to say, this familiarity cannot be conveyed in a page-and-a-half worth of introduction.

    If this is the case authors will move to application justifications for the sake of efficiency, regardless of actual motivation and independently of any biases in the refereeing (or the collapse of western civilization).

    In terms of (2), Oded has failed to make the case that the previous balance was preferable to the current one. Let me remark that, just as there is a risk of too many applications to few explanatory research papers, the reverse unbalance would be equally detrimental to computer science.

  55. @Alex,
    It seems that your claims (which I don’t agree with) actually strengthens Oded’s claims: since it is “harder to explain the intellectual motivations” over the “applicative motivations”, the balance is naturally swayed to the applicative side (*both* in the content, and in the way things are presented). And as the competition gets tougher, naturally, the movement against intellectual values gets stronger, because of your explanation.
    Thus, if we agree that a total unbalance between the intellectual and instrumental aspects is harmful, your claims justify an increase in conscious attention to counter-balance against over instrumentalism in ToC.

    Also, you have presented a possibly new explanation to why there is indeed a movement against intellectual valued works in ToC. Which supports the contentions which Oded sought to prove in the first place.

  56. I cannot resist addressing Alex’s last comment.
    I find that comment fair, although I disagree with it.

    Most importantly, I strongly disagree with Alex’s main point
    (expressed in his 2nd and 3rd paragraphs). But articulating
    my position is for a position paper, not for a note.
    Let me just make a few brief comments:
    (1) applicability is also determined by subjective judgment;
    (2) the fact that something is simpler does not mean one should do it
    (or the fact that something is more difficult does not mean it should be abandoned); and (3) I did hint to these considerations in my essay…

    Regarding Alex’s last paragraph, I agree to his claim that I did not provide any argument as to why the previous balance was better. But I did say explicitly that a bias against instrumental papers would be bad too. The question (*) is what is the right balance. So I think Alex and I agree on all these issues, up to the point of question (*).

    It is not clear how to resolve (*), in the sense of reaching a consensus on it or making arguments about it. Note that we are unable to agree even on simpler issues such as what is the INT-vs-INS dichotomy and whether the balance regarding it has changed at all during the last three decades…

    I cannot hope to convince anybody about the importance of promoting INT, nor did I intend to do that. I can only try to articulate my view as to why INT is crucial for a field of Science (and the same holds for INS!); maybe I’ll do that some day (i.e., write a short essay re this). The point of articulating is triggering thought, appealing and helping those who are similarly oriented, not necessarily convincing (i.e., changing opinions of people from one disposition to an opposite one). Anyhow, I tried to articulate some issues (but not all issues…), and I did not try to convince those who think differently.


  57. Clarification regarding my 3rd paragraph:

    “Regarding Alex’s last paragraph, I agree to his claim that I did not provide any argument as to why the previous balance was better. But I did say explicitly that a bias against instrumental papers would be bad too. The question (*) is what is the right balance. So I think Alex and I agree on all these issues, up to the point of question (*). ”

    What I meant in “all issues” is all issues raised by Alex in the said paragraph (i.e., that it is a question of balance and that I did not articulate why I think the prior balance was better). Recall that we strongly disagree on what he said in prior paragraphs.

    Also, seems that while I was typing “TOC researcher” was articulating some of the things I meant to say better…

  58. @ToC,

    What my argument shows is that simply scoring the presentation in terms of explanatory curiosity is not enough to prove a bias from the community. It could well simply prove a bias for that type of presentation on the side of the researcher.

    From my experience, most people study the things they care about and pursue problems in their area of competence, with only minor attention paid to presentation trends, and this last is usually reflected chiefly in the selection of the venue in which to publish.

    Granted, there have always been people who will do anything to game the system, and they wll avail themselves of every possible shortcut, but fortunately they seem to be a minority.

    Lastly for this discussion to be useful it has to be grounded in reality. First there is no “movement against intellectual values” unless somebody forgot to invite me to the meetings and, second, you and Oded must do without the the loaded term “intellectual values”. Use of non-neutral terms is unscientific.

    You see, it is not possible to have a rational, unbiased discussion between options A and B if we prelabel them, misleadingly “fantastic option A” and “second rate plan B”.

    There is no justification to equate curiosity with “intellectual value”. Explanatory curiosity is but one form of valuable intellectual acitivity. For example, finding a cure for cancer is both a humanitarian and an intellectual pursuit of the first order.

    Lastly, and completely anecdotically, in my opinion the community has swayed over the years from very applied (in the pre-NP days), to highly curiosity driven and back to somewhat more applied. The jury is still out on which of these eras will prove scientifically and intellectually* more valuable.

    * Intellectually in the general sense of the word, not Oded’s,

  59. Alex:

    Oded’s data do not show that there the fraction of applicability-motivated papers increases over the years at the expense of curiosity-motivated papers. What the data show is that the fraction of applicability-motivated papers remains the same over the years while the fraction of curiosity-motivated papers decreases – what increases is the fraction of papers that do not bother motivating their work at all.

    Thus, there is no point debating about what is the proper balance between applicability-driven and curiosity-driven papers, because this balance did not actually change.

    If I understand Oded’s essay correctly, his conclusion from the data is that over the years, people feel less inclined to write “curiosity motivations” in their papers, and prefer writing no motivation at all if curiosity is their only way of justifying the paper. At the same time, they do not feel less inclined to write “applicability motivations”.

    Oded claims that this phenomena means that the importance that curiosity as is given a value in our field is in decline over the years, and this is the problem he points out.

  60. In my last comment, the last sentence should be “the importance that curiosity is given as a value”

  61. Alex, you confuse “intellectual activity” with “intellectual values”, whereas my very first comments in the essay explicitly make this distinction, and does so several times. Ditto wrt the claim that the term is not “neutral”. Again, I refer you to the definitions! (*)

    I do agree that the data only asserts, under some assumptions, that there is a decline in the explicit statement of intellectual motivations. Using add’l assumptions, I infer that this represent a decline in the status of intellectual values. All this is explicit in my essay, which also provides a discussion of the various assumptions including the one supporting the latter inference. It is fair to say that you dispute the assumptions and are unconvinced by my arguments, but you cannot just ignore all of that.

    *) In particular “intellectual values” are not associated with curiosity (and a few other things), but are rather defined that way. Thus, in every place where you see “intellectual values” you may read specific values such as curiosity, study, and desire to understand. The essay distinguishes these important values from another cluster of *important* values called “instrumental” (aimed at measurable progress, applicability, etc). Lastly, I claim that my choice of these terms/labels is quite reasonable and consistent with the common understanding of these terms.

  62. Alex, we did send you an invitation. You should check your spam filter.

  63. Oded, I wonder if something is lost in translation. I cannot find academic references for “Intellectual values” outside from one specific author (Deanna Kuhn). The more common term in English is “curiosity-driven research”, which in my opinion accurately describes the concept without prejudging it.

    Alternatively may I suggest switching to my terminology: “idle mussings” and “useful research” which are also consistent with the common understanding of those terms. Here’s the revised abstract:

    A new informal study by Prof. Goldreich has found that the number of papers on idle mussings has decreased notably over the last two decades. Their place was taken by papers on useful research. The research community will continue to track this trend.

    p.s. this last part was with tongue firmly in cheek.

  64. Alex, I really don’t understand why I deserve this attitude from you. Needless to say, the suggestions that you make in the 2nd paragraph are far more contrived and/or “painted” than the terms that I used.

    The suggestion made in the 1st paragraph is indeed fair, but it does not capture well what I want to capture; please take a look in the beginning of my introduction. I wish to talk of values, not merely motivations. Indeed, in my opinion, values are reflected by the motivations (both the ones stated explicitly and those that can be inferred), but they are not identical. Again, I wish to refer to a cluster of values, one being curiosity-driven”, but others (which I view as no lesser) are study and quest of understanding. You may say all three are related, but (as in many other cases where I insist that related things are neither identical nor reducible to one another) I maintain that the latter two are not reducible to the former. Now, I take this cluster-of-values (not activities!) and label it “intellectual values” (not activities), which seems quite fair to me. In any case, it is but a name, and what counts is how it is defined. Of course, the name is not arbitrary, it is supposed to reflect the definition. This is all standard stuff; they way we do things, so I really don’t understand the fuss.

    I assume one may find a different name/term (i.e.,other than “intellectual values”), but if I am asked to replace then I’d insist that it is better than the one used. In particular, it should reflect the cluster rather than mocks it… However, I must say that I’m not sure that I’ll do such a replacement *at this point*: As I said several times, I’ve spent a month work on this essay, and I’m quite exhausted, let alone that I’m pissed at much of what has been taking place on this blog ever since. I think that at this point a fair minimum prerequisite for anybody asking me to change my terminology *to a better one* (not for a worse one!) is that this person has invested at least a full week’s work on such a project.


  65. @Alex,

    First, by “a movement against…” I meant only “a move against intellectual values” (as in a transition from one state to the another), and *not* a “movement” (as in the social sense), obviously.

    Second, your arguments and examples about the “loaded” term “intellectual” are actually useful.
    For instance, finding a cure for cancer might involve a lot of admirable intellectual effort, but is certainly not an intellectual goal in itself. The argument for this is simple: one can, in principle, find a cure for cancer “by chance”, without any clue to why it works. So the problem itself is completely instrumental. It has no intellectual value (under this extreme assumptions, of course). On the other hand, the development of the theory of NP-completeness, for instance, has, as far as I understand it, an intellectual value in itself.
    Since, I believe that close to 100% of even “ivory-tower intellectuals” would agree that finding a cure to cancer is among the most valuable pursuits of humanity (while they might disagree whether investigating the theory of NP-completeness is so highly valuable), we now can understand (at least I hope so) that the distinction between the intellectual and the instrumental is not only quite clear, it’s also not really biased, or loaded.

  66. “ToC Researcher” is again more on the point than me.
    I just wish to thank him/her! Oded

  67. As usual, I’m joining the blog discussion too late. But it strikes me that the discussion here is blurring two largely orthogonal issues in Oded’s essay.

    1) The classification of research into intellectual and instrumental and the appropriate balance between these.

    I can understand that there are a lot of strong feelings and disagreement about this topic, but it only plays a significant role in the *introduction* to Oded’s essay, and is relatively insignificant in his findings and recommendations.

    2) The extent to which the TOC community continues to hold and express various kinds of values as motivation for their work.

    In his study (see Table 1), Oded found that the extent to which authors are explicitly expressing (intellectual) values as motivation has declined dramatically from STOC 83 to STOC 07. His data does *not* show a shift from intellectual values to instrumental values, so how he classified papers between these two categories is not so important. Instead, he finds that authors are failing to explicitly state motivating values at all (the shift has been from “explicit” to “implicit”). He expresses concern that this shift reflects a decline in values actually held (with the motivation coming more from competition & “symbolic capital”) and may lead to further decline in the future. So he makes recommendations for how to restore the values of the community.

    I think these findings and recommendations should be of interest and worthy of discussion regardless of how one feels about the balance between different types of TOC research. (Oded’s essay is clearly more focused on what is happening on what he considers to be the “intellectually motivated” side of TOC research, but a decline in expressed or held motivating values should be of concern for all sides of TOC research.)


  68. TOC Researcher said: (about the value of intellectual activity) ““and in most non-Western societies, there is no decline, because there was no value to it from the start””

    I think I missed the announcement that India, China, Japan, Egypt, Iran, Indo-China, and the rest of Africa and Asia and the ancient civilizations that flourished there have now been declared “Western societies”.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s