The nature of morphological productivity has been a topic of debate in Dutch morphology circles since Schultink’s famous article on this topic ((Schultink 1961)), and his dissertation(Schultink 1962). Schultink reserved the notion of morphological productivity for cases of unintentionally coined new words, and considered other types of formation of new words a case of creativity. His famous definition, often cited in the international literature on morphology reads as follows: 'Productivity as a morphological phenomenon is the possibility for language users to coin, unintentionally, an in principle uncountable number of new words by means of the morphological process that underlies the form-meaning correspondence in some words they know' ("Onder produktiviteit als morfologisch fenomeen verstaan we dan de voor taalgebruikers bestaande mogelijkheid onopzettelijk door middel van het morfologisch procedé dat aan de vorm-betekenis-correspondentie van sommige hun bekende woorden ten grondslag ligt een in principe niet telbaar aantal nieuwe formaties toe te voegen" (Schultink 1962: 113). This distinction between productivity and creativity is also defended in (Van Marle 1985). However, other linguists have pointed out that it is not easy to make a principled distinction between unintentional and intentional language use (Bauer 2001).

The actual use of in principle productive processes differs enormously. Various factors play a role, such as the semantic usefulness of the morphological category, the number of available base words, style and text type, the competition with existing words that already have that meaning, and the competition of other morphological categories with similar meanings. This may be phrased as follows: not all possible complex words (as defined by the set of morphological schemata) are also probable complex words (Van Santen 1992).

Another complication is that a word formation process may only be productive for certain subcategories. For instance, the formation of denominal nouns in –er such as wetenschapperscientist and Amsterdammerinhabitant of Amsterdam is not productive across the board: it is productive for coining inhabitant names, but a noun like godsdienst-erreligion-erreligious person is odd.

The way in which the degree of productivity of a morphological pattern can be determined is a matter of debate. It is not sufficient to only count the number of word types of a category, and assume that a high number of word types indicates a high degree of productivity, because unproductive categories may also be instantiated by a high number of word types. Therefore, (Baayen 1989) proposed to measure the productivity of word formation processes by counting the number of hapaxes (word types with frequency 1) in a large corpus. This way of measuring productivity is based on the idea that a productive process will lead to a lot of hapaxes, whereas processes with a low degree of productivity will not result in a lot of such low frequency word types.

The issue of productivity, with examples from Dutch, is discussed in detail in (Booij 2017: chapter 6) (cf. also (Barðdal 2008: Ch. 2). Baayen has published a number of studies on measuring productivity of Dutch word formation processes (Baayen 1989), (Baayen 1993), (Baayen 1994). Clear cases of unproductive word formation in Dutch with many types are verbs in –eren and –elen such as bibberento shiver and babbelento chatter. They are dealt with in (Booij and Audring 2018) who argue that the morphological schemas that characterize these sets of verbs have the function of motivation: they express that the meaning of such verbs is not completely arbitrary.