I think considering a genome as a recipe is a decent analogy. However, not all parts of the genome are under the same kind or degree of regulation. In fact, it's unlikely that the regulation on one portion of the genome even resembles that of any other due to the staggering complexity of potential regulation. So some genes, and therefore traits, will be more sensitive to the environment than others. As an example, after fertilization a zygote is going to start dividing no matter what environmental signals are out there. I would imagine that a core trait such as facial features would follow its genetic recipe very closely, and only be influenced by the environment under very extreme circumstances.

Sounds like a discussion regarding semantics, whether to use "recipe" as it is used in cooking or to use "blueprint" perhaps as it is used in building. But in this discussion both are just metaphors. Is cooking food an easier metaphor through which to view development because it is easier to see the variation when one recipe is used to cook a dish multiple times yet there are easily recognized differences in the results, as a result of small differences in, say, the temperature of the oven or the amount of mixing of ingredients prior to cooking or how it is presented for consumption? Do we think of blueprints and buildings having less variation from one building to the next? Or is that just a cultural/historical difference? I think either one is fine, neither is better or worse. The important thing to remember is variation in the genome among individuals, some affect contributed by the environment and genetic selection on that variation over generations. Is the genome acting like a recipe or a blueprint? These are just words. It depends on what these words mean to the reader.

I agree with Rock that the question is to some extent semantic. Personally, I think the concept of recipe is closer to what really happens in development. We all have body hair, some more than others. There is no blueprint of where each hair has to be placed, but a series of cell communication events with determine the rough area to be covered by hair, the spacing of hairs, their thickness, etc.

In contrast, the head of a toy doll is made according to a blueprint. There is a cast with the wholes where the hair will be placed.

Still, a main difference between a recipe and development is that the second happens within the context of an egg that already carries information. The main body axes are determined before the genome of the embryo is activated. So development is close to a genetic recipe with some embryonic pre-cooking involved.

I do not fully agree with your statement:

"if it should happen to be fact that the information that governs those parts of development that lead to facial structure indeed originate within the genome, then it must be that some portion of the genome acts as a blueprint and not a recipe"

It is more likely that it is not some portion in particular, it could be several of them and their interaction. Going back to the metaphor of the recipe, if my bread is spongy (as it should be), it should be attributed to having got right several conditions: proportion of yeast, fermentation time, baking temperature and baking time. Moreover, changing just one of these factors will affect not only sponginess, but also size, crust thickness and hardness, moisture, etc.

Mutations in EGFR, to mention just an example, can lead to defects in cell proliferation, survival, apoptosis, neurogenesis, cell fate and many more. Even more, EGFR acts throughout development and adult life, in slightly different, and sometimes completely opposite functions. So we cannot say that EGFR is part of a blueprint for something. It is more like a sentence in the recipe that will affect several of the final features.

The difference between blueprint and recipe could not be more clear - a cake is the product of a recipe (where no portion is specified in terms of place and type, since all parts are identical and no part of the cake is different from any other part of the cake) while with blueprint the placement of parts is exact with respect to place and type.

The important point is that a full accounting of developmental processes must provide a clear, causal and understandable (such as in terms of man coming into possession of such knowledge as will leverage itself over those developmental processes - especially by construction of physical models that exhibit similar processes - machines that develop) description of such processes.  Else, our understanding of developmental processes amounts to mysticism.

It is not enough to say that the mechanism of control over physical development is distributed; such fact does not make the process inherently unintelligible.  We accept that there must be a system of causal relationships that provide the knobs that when turned (alterations to regulation of a gene) yield changes to phenotype, and that such settings are the reason for uncanny similarity between generations, even between two persons who look similar but are otherwise unrelated.

Please explain to me how the recipe/blueprint could be otherwise.

@Maximo - Two points.

1. It is no limitation that a portion of the genome (such as would govern exhibition of a phenotype in an individual) occurs in distributed fashion; a portion is a portion, contiguous or not.

2. Further, should the removal of some portion (even a dis-contiguous portion) of the genome yield a predictable removal of a phenotype (and the introduction of that portion into the genome of an individual yields expression of the phenotype), then at least that portion of the genome can be more likened to a blueprint - a (pheno)type in a particular location appears in response to the transcription of a specific portion of genome.