Many of my students were a little confused in my class on "modern Smithians" when we talked about Romer and Krugman's work on growth and trade theory, respectively, as being two of the preeminent examples of modern Smithian economics. They didn't see division of labor anywhere in either of those works - at least not obviously so.
We just got into neoclassicism yesterday, after having spent two classes on marginalism (neoclassicism, for these purposes, effectively covered the marginalists after the marginal revolution). In the intervening weeks they learned about how Smithian productive efficiencies quickly gave way to diminishing returns in Malthus and Ricardo. Both of these economists used diminishing returns because it was plausible, but also because it provided a stable solution to the problems they were considering: population and rent.
We left that aside for a couple weeks working through Mill and Marx, but then came back to these assumptions when we started into the marginalists. Mixing together primary sources, some standard intermediate micro problems, and out history of thought text the importance of diminishing returns really started to crystalize for them.
Then in steps Marshall and others who applied these tools to problems of interest. Marshall tells one particular story using the language of marginalism, but it's a story about how some industrial districts seem to be characterized by increasing returns. Assume spillovers. Assume nice pinned down firm behavior with the activity of other firms bleeding over into each others' cost functions. Now we have a case that bucks the trend of diminishing returns we've been seeing week after week after week. Smith was so nice and positive - Malthus and Ricardo were both downers. And here comes Marshall with a happy result again that is very reminiscent of Smith.
Where have we heard something like this before? I asked.
And then several students' eyes lit up. That's why Krugman and Romer made a big splash. It seemed so different from Smith at the time because the story was in a different language, but after trudging through Malthus, Ricardo, Mill, and the marginalists and understanding the importance of the increasing/diminishing returns divide, Marshall's spillovers seemed refreshingly Smithian. It's not just increasing returns either, it's the idea that productivity goes up with densely populated areas too.
Anyway, this connect the dots exercise is obvious for economists. But if you've read a little Smith on the division of labor and you don't have the context of the next hundred years of economic thought it's not exactly clear why I said what I said earlier in the course.
I am not one of those people that says that all undergrads should take history of thought. I think that's an overstatement and unnecessary. In a progressive science (NOT Whiggish... just progressive) we shouldn't have to study the history of thought to be good scientists. But, the fact remains that if you do know history of thought you have a much richer and more contextualized understanding of the science as it's practiced today.