The term compositionality in the domain of tense, mood and aspect prevalent in the current study of language, logic and computation is clearly to be understood as derived from the notion of function composition. Linguistics and logic were closely connected in traditional grammar but their modernization resulted in a clear break: logic said farewell to Aristotelian logic and the same happened for quite different reasons in linguistics which focussed on diachronic research in which the role of logic in language became marginal as it remained in structuralism.
The focus of diachronic research was increasingly on comparing older stages of different languages with one another. Hence scholars could not escape from dealing with the notion of translational equivalence: comparing languages means that something expressed in one language is equivalent up to a certain point to something expressed in another language. Up to a certain point, indeed. German scholars writing about the aspectual system of Slavic languages—there were many of them—were well aware of the fact that what was mainly expressed morphologically in the form of aspectual pairs in Slavic languages could easily be translated into German. Yet two factors prevented them from drawing the right conclusion. Firstly, they did not have a syntactic framework in which they could embed the intuitively felt semantic equivalence; secondly they were hampered by the assumption that each language has its unique features distinguishing it from other languages. They did not have the tools for dealing with compositionality. These tools were available for me when I wrote my PhD at the end of the sixties.
I will try to explain which factors did prevent scholars in the first half of the twentieth century from drawing the right conclusions from what they observed and which factors enabled me to draw them. After more than fifty years I will return to Poutsma (1926) and Jacobsohn (1933) in order to point out why I consider them my predecessors and why they could not do what I could do with the help of syntactic trees: to see aspect of a sentence as being composed of smaller semantic units. I will also indicate why 50 years were necessary to really improve on that.