I can accept the notion of threshold concepts (TC) as a catchall, shorthand for concepts that typically prove troublesome to a significant number of students during a typical learning curve. Many students hit a learning wall during their development. The experience of many learners and teachers points to liminal spaces through those walls that once crossed offer the learner deeper insights into that subject in erstwhile unimaginable ways.
An example in English education in Japan is the BE verb. Most students begin equating it with the Japanese topic marker. Only when they ‘see’ its grammatical function do they create expressive and dynamic English sentences. This cognitive leap often does not happen until the mid-intermediate levels and beyond.
A major question we need to ask is if these TC are a function of the inherent complexity in a subject (i.e. most learners are likely to experience trouble with them) or are a function of instructional design (ID) (i.e. due to the typical presentation of information). Research points to the latter, yet many writers simply assume the former. Marton and Pang (2006) show how better ID can overcome difficulties. As Rowbottom explains, “concepts come in complex and atomic versions” (Rowbottom, 2007, p. 266). When I show an atomic analysis of the BE verb against the Japanese sentence structure, elementary students usually do not hit any walls later.
Without a clear theory of TC, a facile application of them is akin to people today using the term race in the 19th-century meaning: useful but flawed.
Marton, F., & Pang, M. F. (2006). On Some Necessary Conditions of Learning. Journal of the Learning Sciences, 15(2), 37–41. doi:10.1207/s15327809jls1502
Rowbottom, D. P. (2007). Demystifying threshold concepts. Journal of Philosophy of Education, 41(2), 263–270. doi:10.1111/j.1467-9752.2007.00554.x