Many of the chemistries studied by chemists as being relevant to the origins of life produce a combinatorial explosion of products. Researchers in the origins of life have traditionally regarded such combinatorial explosions as a problem that needs to be overcome, but due to recent technological advances there is an increasing interest in understanding these combinatorially complex chemistries as complex dynamical systems in their own right. An important question is whether a sufficiently large and complex reaction network can, in some quantitative sense, behave fundamentally differently from what is possible in the clean chemistries more usually studied by chemists. Here I argue for an affirmative answer to this question, with the aid of a simple toy example. The example shows the existence of a phase transition, leading to a threshold effect of a kind that cannot occur in a small reaction network. This threshold is closely re- lated to Eigens error threshold. However, while the error threshold occurs in a very structured chemistry (template replication), this transition occurs in a very unstructured chemistry more closely resembling Kauffmans autocatalytic set model. The significance of this result lies not in the scenario presented, but in showing that such threshold phenomena can exist in messy chemistries at all.