The project focuses on shifts in the productivity of two clusters of constructions in Belgian andNetherlandic Dutch over the last two centuries. The linguistic phenomena under investigation are:(i) patterns such as een schat van een kind (lit. a treasure of a child ‘a sweet child’) and een draakvan een wedstrijd (lit. a dragon of a game ‘a boring game’), which express a kind of metaphoricalcomparison (‘a child like a treasure’), and (ii) patterns such as Hij piekerde zich dood (lit. Heworried himself dead) and Hij zweette zich een ongeluk (lit. He sweated himself an accident) whichdo not describe a situation in which someone actually dies or has an accident, but rather expressthat the activity denoted by the verb is carried out to excess (i.e. ‘He worried/sweated verymuch’). Both constructions are expressive in meaning, and both display an interesting mix of“fixedness” and “openness”. For instance, while we can both say Hij lachte zich een ongeluk andHij werkte zich een ongeluk (‘He laughed/worked very hard’), the possibilities of een hoedje (‘alittle hat’) are much more constrained: Hij lachte zich een hoedje is fine but Hij werkte zich eenhoedje is very odd. There are indications that such constraints have relaxed over time, i.e. that theconstructions are more productive now than they were in 19th and early 20th C Dutch. Theinvestigation of such shifts in productivity can shed important new light on the recent theoreticalconcept of “constructionalization”.