A couple weeks ago I posted an opinion piece more-or-less on how I think biologists should work on a more precise language, one example being the language of dominance/recessiveness in genetics. Douglas Allchin is skeptical as to the usefulness of such language and believes we should “dissolve dominance.” Most of the comments agreed with me, but Graves asked me to clarify a bit.
While I was intentionally provocative in my post, I do agree with its sentiment. The importance of precise language has been drilled into my head by multiple teachers and I have witnessed the wonders strong editing can do to a clunky sentence. (I am aware my writing on the blog can be imprecise but my posts only go through a minimal editing process.) While those teachers were from the English and history disciplines, I think an even stronger case can be made for precision in science with its jargon-laden language.
While reading Razib Khan’s post about the reconciliation of Mendelians and biometricians by RA Fisher, I noticed he briefly discusses the genetics of cystic fibrosis with dominance/recessive language. With this I realized that one of the best pro-arguments for dominance language is its generalizability (new word!). Without having to read about cystic fibrosis in depth, I was to some extent able to understand its genetics.
Consider a rare recessively expressed allele, one for cystic fibrosis. As it is rare in a population in almost all cases where the offspring are homozygotes for the disease causing allele, both parents will be heterozygotes. They will not express the disease because of its recessive character.
Immediately I understand the two alleles in some vague sense – I kind of understand their relation to each other but I have no idea as to the molecular mechanisms of that relation. That’s the key. This “debate” is between two extremes: generalizable but vague vs. precise but very specific and specialized. Biologists have to decide on which side the preferred language resides.
As for Graves’ question specifically, what Allchin prescribes isn’t an alternative to dominance in a strict sense as he thinks there is no use in replacing it with less-loaded terms. Instead, he thinks we should just describe what is actually happening. As I mentioned, his paper is available for free here (PDF). If you don’t have the time to read the whole thing, I suggest at least reading “Section VI. Recasting Dominance” (page 13). Here he provides examples of how dominant/recessive fails to explain phenomena (such as blood type and sickle cell anemia) and how we benefit from a more nuanced understanding of them. (He also notes how Mendel did not use dominance language in his own work.)
I have no specific recommendations myself as I am not a teacher. Could genetics be taught without dominance?, i.e., could high school/freshmen students understand the molecular mechanics at work? I believe that teachers should at least tell students that dominance is more or less a theoretical construct that allows us to understand at some vague level how genetics works (and by vague I mean I’m skeptical of its usefulness at all) and that expression discrepancies between alleles can be explained by many many mechanisms. As I mentioned in the initial post, I found my professor’s story about how Mendel’s peas actually worked enlightening. (I’m fascinated by science that investigates historical problems and sheds new light on topics of which we thought we knew all there was to know.)
For myself, I don’t see much use in dominance (unless someone shows me otherwise – I’m open to changing my mind), but as a teaching tool I can see it being useful. As to how, I’m not sure. I think it’s hard for me to imagine learning genetics without dominance as it is how I was taught. I think even replacing dominance/recessiveness with functional/afunctional would help describe many cases. As for blood type, while co-dominance for A/B and dominance for A/B over O is useful, perhaps they could be better described as co-expressed (AB) and non-present (O). Less generalizable maybe, but a better descriptor perhaps.
Again, I would recommend reading the last few pages of Allchin’s paper for examples. I would love to hear some critiques if anyone has some!