I'm not sure if Science is phrenology by another name, but they certainly look similar at your level of detail:
a) Part of the task under study is definitely linguistic
and
b) No effort has been made to separate this linguistic part from the rest of the task
so
c) The study has not produced evidence because of a validity threat: namely the confounding factor that the task has been presented in a linguistic form.
I wouldn't want the terminology to get in the way of the original point: the task has been phrased in textual form, areas of the brain used in text recognition lit up, the researchers concluded that programming was the same as language skills. Their conclusion was bogus because presenting a non-programming task to the participants would have provoked the same response if it was done in written form. Obviously this would be impossible to fix in the study design.... without replicating the results on a non-textual experiment, such as a graphical language.