Plato’s weekend suggestions, episode 51

readingsHere it is, our regular Friday diet of suggested readings for the weekend:

The American Philosophical Association has just published its new code of professional conduct. Problem is, among other things it calls for an end to sarcasm

Are there six basic types of philosophers?

This is the real problem with GMOs, and why I want labels. (Yes, I’m aware of the responses to the NYT article, extended commentary coming on Monday.)

The new book of snobs, an ever-popular attitude among Homo sapiens.

Toward a new (evolutionary) philosophy of science?

The internet was supposed to herald the golden age of intelligent information. Didn’t happen.

The real problem of consciousness (since the hard one doesn’t exist…).

Is there a problem with bibliotherapy?

77 thoughts on “Plato’s weekend suggestions, episode 51

  1. dbholmes

    Hi Coel, I should add that all the author says about the first graph is that “the long-term yield trend for both areas is up.” It didn’t specify that they were up in equal amounts. This is correct.

    Perhaps you mistook the commentary below it (which talks about “keeping pace”) as being about the first graph? It is about the second graph, which is clearly close. Indeed the author could throw out the first graph and keep the other two. They don’t help GMO advocacy at all, and so calls into question why the author’d have to cheat with the first (rather than making a common mistake).


  2. dbholmes

    Hi Coel, now I cross posted over two of yours!

    Point taken on language within the over all text (which is different than that seen in the graph area itself), though I think some of the quotes you gave relied on the other graphs as well.

    “I’m not disagreeing with you on the wider points and the merits of GMOs. And I think that you’re not disagreeing with me on the narrow point of the merits of that graph. So I don’t think we’re disagreeing at all. ”

    Yes, good.

    “That must be some sort of first for this blog”

    I disagree! 🙂

    Actually just a short while back I was arguing in support of your and DM’s position. I forgot on what now… but I did! It happens!


  3. Massimo Post author


    We are just going to agree to disagree on the value of that graph. Of course if the raw data is in absolute terms then the slope is expressed as a function of absolute data. But when you say that it is much more normal to compare things in terms of percentages I don’t think you are reflecting any consensus at all. It depends on what one is trying to do. You may disagree that what the NYT author was trying to do is the best / most informative thing to do, but you have no argument that things “nearly always” need to be expressed as percentages. What would be an exception, by the way? In this case, if the author had done as you suggest I can easily the reverse criticism: he didn’t take into account that the baseline are different! Moreover, a coincidence? Hmm… And your paranoia about cheery picking and pre-determined conclusions is, I think, just that, paranoia. Finally, you keep ignoring the fact that other sources, also cited in the article, confirm the conclusion that there are no major differences in yield between the two cases, quite regardless of how the data is plotted.

    Liked by 1 person

  4. Massimo Post author


    Also, I’m curious, how do you explain the exact same result, in terms of slopes, of the second graph, where the baselines are identical? Another coincidence?


  5. Coel

    Hi Massimo,

    Also, I’m curious, how do you explain the exact same result, in terms of slopes, of the second graph, where the baselines are identical? Another coincidence?

    Nope, the second graph seems fair. And as per my comments to dbholmes, my complaint really is a narrow one about that headline graph. The wider issue is a rather wider issue.

    As for why those two graphs are inconsistent (in terms of percentage improvement), well all sorts of things would be different between the two situations, and I don’t claim to know what explains the difference in increase in yield.


  6. Robin Herbert

    Two trends that are parallel in absolute data will always converge in terms of percentage increase or decrease. If it is a linear trend in the absolute data, as the first graph is, the percentage increases will become smaller and converge with each other – the two trends are both asymptotically approaching the same percentage trend increase.

    What that means depends upon the underlying thing being measured. If it were increase in disk used, for example, if you have two 10 GB disks and in one there is 3 GB in use and the second has 5GB in use and the are both growing at 100 MB per day, the disk first disk has a greater percentage increase than the second but it would not be true to say that their is a difference in the data growth rate.

    With hectogram yield per hectare, I am no expert, but I wouldn’t have thought that it was very significant if the difference in percentage growth is simply an artifact of the starting values and when the percentage growth in both is decreasing and approaching the same asymptote.. It means the Canadian data, on this trend will never catch up with the Western Europe data.

    There are other points, for example only part of the Canadian data has the GMO crop and there does not seem to be any change in the trend. And, as Massimo points out the corn crop yield, coming off the same base has just the same trend, both in absolute and percentage terms.

    All in all I think the data bears out the claims of the article.

    Liked by 2 people

  7. Robin Herbert

    For the percentage increase of the Canadian rapeseed crop yield to be parallel with the Western Europe percentage increase, the increase in the Canadian crop yield would be getting smaller each year, while the Western Europe percentage increase stayed the same year to year. So there is really no inconsistency between the graphs.


  8. marc levesque


    I don’t see any of the problems you mention.

    Looking at the graph I see that both lines, Europe and Canada, show an increase of about 2000 hectograms per hectare every 5 years, and that that rise in yield is independent of whether GMO varieties are planted or not and also independent of the number of hectograms produced per hectare (which is interesting in itself but besides the point). So that fits in well with the one of the author’s main theses that “genetic modification in the United States and Canada has not accelerated increases in crop yields”.


  9. marc levesque


    I enjoyed the selection. This hit me from Philosophy Beyond the Academy “He aims not to eliminate all puzzlement but to leave us puzzled about the right things”. That sounds about right to me and of course it still leaves open what are the right things to be puzzled about.

    (and last week I really enjoyed the the piece on disability and how being used as a model of resilience among other things can also be demeaning or abusive)

    By the way, I like your blog’s new title, and congrats on your blog award.

    Liked by 1 person

  10. marc levesque


    If you compare crop yield rise as a percentage of base yield, comparing Canadian non GMO rape seed for 1985 to 1995 and Canadian GMO rape seed for 1995 to 2005, you probably get a bit higher percentage rise for the non GMO variety. But I don’t mean that as an argument in favor non GMO varieties, only as a contrast to your point.

    J erome,

    “As for the more direct comparison of rape seed yield, it would be relevant to know what the rape seed GMO variety was for. As far as i know there is no GMO modification to increase yield so why compare relative yields?”

    As far as I know bio-tech manufactures say that most GM modifications are related to yield. For example if bugs are eating your corn you can buy a GM corn with a bug toxin in it’s genetic code so your crop yield won’t suffer as much from the insect, or if weeds are the main reason your corn crop yield has stagnated, and you can’t use more herbicides because the amount you’re using is already starting to affect your crop badly, then you can buy a herbicide resistant GM corn and spray more herbicide to get the weeds down and raise your yield (I’m pretty sure GM canola has the resistance to herbicide modification too).

    Liked by 1 person

  11. Coel

    Hi Massimo,

    But when you say that it is much more normal to compare things in terms of percentages I don’t think you are reflecting any consensus at all. It depends on what one is trying to do.

    I agree, it depends what you are trying to do. If one presents that plot in absolute units, then the difference in initial starting point becomes a dominant feature of the plot. So, is the text mainly about the difference in the starting points? Well no, not really, it hardly discusses that.

    What the article does discuss is the change over the time period. That’s why it would be far more usual to normalise to the initial starting point (thus removing the initial difference from the plot) and plot percentage changes from there.


  12. J erome

    Marc, let’s be clear about this. Current GMO modifications do not directly increase yield, they increase pest resistance or herbicide resistance. This allows you to use less pesticide or more herbicide respectively. There may be an indirect effect on yield as you mentioned. In the case of Canola the GMO variety allows the use of more herbicide allowing a change in tilling practice which should reduce soil erosion. It likely will not increase yield.


  13. marc levesque

    J erome,

    “As far as i know there is no GMO modification to increase yield”

    My point wasn’t direct or indirect influences, types of yield increases, particular crops, or any of the myriad of farming techniques or how they can by combined in multiple ways :

    “At Syngenta, we aim to contribute to food security and sustainability in agriculture by helping farmers improve yields …”


Comments are closed.