Metaphor fight! Shenk and Dobbs Square Off!


In response to this month's Atlantic feature "The Science of Success," by David Dobbs, which I admired, I invited Dobbs to engage in short back-and-forth over one particular gripe I had. He graciously accepted. Children, avert your eyes. This is literary brawling the likes of which haven't been seen since Norman Mailer head-butted Gore Vidal.

Shenk alights from behind a doorway with his first jab:

Congratulations on your beautifully rendered "Orchid" piece. You do a superb job of illustrating the notion that the same gene can yield very different results in different circumstances. I particularly admire the way you end the piece -- falling back on the essential truth of the parent helping to constantly flip little genetic switches in the child. I consider this piece a landmark step forward in the difficult transition of helping the public understand what genetic expression is all about.

My one not-so-small quibble is that I think you let the metaphor get away from you a bit. While the "orchid" metaphor is a provocative way to illustrate that certain genes or combinations of genes might increase plasticity, the "dandelion" half of the metaphor strongly suggests that "most of us" don't have very much plasticity -- i.e., that the dandelion kids don't have much potential to be either down-and-out or enormously successful. Being familiar with some of your previous work, I don't think that's the message you intend to send. 

Correct me if I'm wrong, but I don't think that you actually believe that science has demonstrated that most of us are destined to a hardy mediocrity. If you are taking that position, I'll respectfully disagree and let's debate that point. 

Your earlier work (which helped to inspire my forthcoming book) suggests that you have a very keen understanding of the extraordinary plasticity built into virtually all of us. I submit that this doesn't contradict the science in your new piece. We can recognize certain extraordinary orchid alleles without rhetorically ghettoizing the other alleles as not-very-plastic dandelion weeds. After all, the studies you cite are presenting population percentages -- they are not showing a clean separation between individuals with or without the alleles. Clearly, as this science marches on, we're going to be stumbling onto specific genes and combinations that seem to have a particular influence in one direction or another. But as we do, I think we need to be careful not to overstate their lessons. We don't want to leave readers with the impression that, without a particular allele, a person is protected from being depressed or barred from having super-talent. 

To sum up, and I know this inconvenient for you, I suggest that you need to drop the "dandelion" half of the metaphor. It's a vivid contrast to the orchid metaphor, but I believe it's too misleading.

Dobbs side-steps, casually finishes his drink, and winds up:

Thanks for the nod and the good questions. 

Your main concern seems to be about plasticity  -- specifically, whether my contrast between so-called orchids and deadlines is meant to suggest that the dandelions are less plastic overall. 

It's a good question, and it lets me make two distinctions that need to be made in a fuller account of this hypothesis. One regards how clean or absolute a distinction we should make between so-called orchid people and so-called dandelions. The other regards whether the plasticity spoken of in this orchid hypothesis is the same plasticity that you and others (including myself) write about as the key to learning and gaining expertise. 

Let's talk about the distinction between orchards and dandelions first, and then what sort of plasticity this orchid hypothesis is mainly concerned with.

Every metaphor has its limits, and one of the limits of the orchid versus dandelions metaphor is that it implies a binary, A or B. division of personality types determined by behavioral gene variants: you're either orchid or dandelion. That's not quite accurate, for there are several genes in question here, and because we each get a mix of variants among them, it would be a rare person that was all orchid, so to speak, or all dandelion. 

To explain: Behavioral geneticists have identified somewhere from 5 to 15 (depending on whom you ask) genes whose variants are important in shaping temperament and behavior. (The serotonin transporter gene, or SERT gene, which is associated with depression, among other things, is the most well known.  But there are several others, such as the MAOA gene, which affects aggression and sociability, and the DRD4 gene,  which affects behaviors on both the attention-distractibility spectrum and sociability in cooperation.) 

For argument's sake, let's say there are 10. In all ten, the 'dandelion' form is the most common, with the orchid forms accounting for about 20 to 35 percent. So for any given one of these genes, you're more likely to have the dandelion variant than the orchid. However, odds being what they are, you are also likely to have the orchid form in at least some of these genes. And since the overall effects on temperamental plasticity are presumed to be multigenic, more orchid genes you have, the more temperamentally malleable and mercurial you will be. In addition, the particular combination of genes in which you have the orchid form will color the nature of your malleability. 

Put that all together and you see that pretty much everyone has some malleable variants in them, with the mix varying quite a bit. I might have four orchid varients among the Big Ten, my wife might have two, and my brother five. So everyone -- if this isn't getting too cute -- some orchid in him and quite a bit of dandelion. So it's not that a person is either plastic or not. The malleability runs along a spectrum, and is a matter of hue as well as intensity. And the consequences of that malleability, of course, depend heavily on experience, context, etc. But the more malleable folks are shaped more dramatically by their experience and react more dramatically, in temperament and behavior, than the less malleable. 

Okay, so that's orchids versus dandelions -- a spectrum issue, not a black-and-white one.

But what exactly is more plastic and reactive in orchids? Is this the same plasticity we talk of when we talk about learning and the mastery of skills and expertise? 

Well -- no. One is temperamental plasticity; the other is cognitive plasticity. There is surely some overlap. But at least by the terms of this orchid or sensitivity hypothesis,  the genetic underpinnings and dynamics of temperamental plasticity are not those of cognitive plasticity. Having the more plastic "S/S" form of the serotonin transporter gene, for instance, will tie your vulnerability to depression more closely to your experience, but it will not make you either smarter or duller, or faster or slower to learn -- especially when the learning involved is primarily cognitive or skill-based, such as learning catalysts, chess, HTML coding, or a good topsin backhand.

So I am not suggesting -- and I don't believe those working up this hypothesis suggest -- that those with so-called dandelion genes are destined to a hearty mediocrity. Rather, the hypothesis asserts that, as with orchids, a dandelion's ultimate endpoint and accomplishments will be determined by a complex mixture of temperament and cognitive and other skills -- but the dandelion's path to wherever will likely be bit steadier and less likely pushed up or down by great or horrible fortune.

Bell rings. Shenk retreats to his corner to tape a small cut. Tosses away gloves, pulls out sword:

Thanks for engaging. Before I dig in again, I want to first repeat that I applaud you for exploring the dimensions of this new hypothesis and working to translate it to public consciousness. When you discuss stuff like multigenic effects and temperamental plasticity, and describe genes as "shaping" rather than causing, and talk about the spectrum of effects, you are conveying some very important nuances about how genes actually work. Even though many scientists understand this stuff, the public has no real clue, and a shocking number of science writers are so far resistant to abandon outdated metaphors and determinist phrases. 

I have to be honest that I'm not entirely sold on all particulars of the nuanced science you're conveying. But that's for another forum, and for time and more studies to clarify. What I want to focus on here is what I see as the only serious problem with your message, which is your metaphor as you are currently using it. And I have a specific suggestion about how to fix what I perceive as the problem. 

You say above that every metaphor has its limits, and I completely agree. Metaphors can only be used to convey a basic essence; on closer inspection, the analogy to the more complex reality never holds. I have no intention of holding you to an unrealistic standard. My problem is with the essence of your metaphor (as you are using it). In your Atlantic piece, the very purpose of the metaphor is to convey that there are two distinct types of kids, the orchids who have much temperamental plasticity and the dandelions who have little. The very first line in the piece is, "Most of us have genes that make us as hardy as dandelions."

What you're acknowledging in your note above is that the orchid kids/dandelion kids distinction is not real; that in reality, there are *orchid genes* and *dandelion genes*, and at the particular mix we each inherit will affect how plastic our temperament is. 

So this is my plea, as you begin expanding this into a book: if you decide to stick with the orchid/dandelion metaphor, don't apply it to people. Instead, apply it only to the genetic ingredients in people. I realize this approach has less literary cachet, and may constrain your evolutionary argument. But it also doesn't build in a dramatic misperception from the get-go.

The other option, of course, is to choose a new metaphor altogether. 

Dobbs reveals spikes on his shoes, takes long leap toward Shenk's abdomen:

I can understand your problems with my message and the metaphor -- and I may (or may not) yet change either the way I use this orchid-dandelion comparison or use something else. That said, I'm not convinced the problems are as serious as you think. I could be wrong about this, of course, and would love to hear what other readers think -- if very many are confused, thrown off, etc., I certainly want to know. 

But, as I say, I see the problems as less serious than you do. 

First, the metaphor is used in the Atlantic piece not primarily to distinguish between stable dandelions and more plastic orchids, but to use that admittedly stark (possibly overly stark?) contrast to replace the conventional one between supposedly stable 'protective-gene' people and fragile or vulnerable (rather than malleable) holders of 'risk' genes.  The reactive and less reactive replace the vulnerable and the resilient. 

(As to the first words of the piece: The passage you quote is actually a sort of deck to the article, added by the editors. It is a bit misleading, and arguably too stark and absolute, and if I had to do it over again, I would have changed the language there so it was less so.  I failed to because it was added very late in the going, and I didn't examine it hard enough amid my concentration on reviewing again the story and some last-minute changes in it. My bad; I can see how it frames the metaphor as you say, very A or B.)

That said, I think most people recognize that decks or editorial synposes, like titles, contain some oversimplification, and that readers need to see the article for the full story. 

More important, I think most people accept contrasting terms such as orchid v dandelion, when used to refer to people, as types that represent either end of a spectrum. That's the case when we talk of extroverts and introverts, for instance; we use those terms all the time, and everyone understands that we're talking about people on one half of spectrum or another. Likewise, I'd guess, with orchids and dandelions. There's obviously a danger if a writer presents these outright as distinct types with no overlap. You seem to feel I've done so; possibly so. I disagree. I think if a writer provides the fuller story behind a contrast like this, most readers understand that the shorthand refers to types at either end, not two mutually exclusive categories with no overlap. 

That said, I'm pondering this and will continue to consider whether the orchid-dandelion scheme is the best way to denote this difference in temperament. A metaphor or distinction like orchid versus dandelion can have great value in giving people a concise image or idea around which to gather a broader and more complex idea; the many, many readers who have reacted strongly and with good understanding of this story -- readers who get its essence quite clearly -- makes me think the orchid-dandelion contrast has that value here. 

At the same time, that strength is a weakness (whoa! meta-alert!) if the metaphor leads people to an inaccurate or wrong idea -- and that has happened with a few readers as well. (Though most of the misunderstanding seems to come from people who have not actually read the article.) Before deciding whether to abandon the orchid-dandelion language, I've love to get a better sense of that balance (that is, how many readers it worked well for and how many it misled) -- and to consider as well how a book-length treatment might overcome any shortcomings.

In any case, your cautions and criticisms rise from healthy concerns, and while I'm not sure we'll meet more than halfway, I value the prod, the critique, and the attention to subtleties of language, whether in title, flap copy, or text. With luck, readers will chime in and give us both a better read on how well the orchid-dandelion distinction works for them.


Sounds fair. Thanks for the dialogue.


Thank you.

Both men collapse on floor. Medics rush in with protein smoothies and the latest issue of Nature Genetics.