Unification can be expressed by writing a variable name, ":=" and a field name, so: $uniVar := uniField If this is all by itself, it's just a unification: FactType( ..., $uniVar := uniField, ... ) It can also be tacked onto a constraint: FactType( ..., $uniVar := uniField == something, ... ) which is semantically the same as FactType( ..., $uniVar := uniField, uniField == something, ... ) But the parser (in 5.4.0) also accepts such a construct: FactType( ..., xField == 42 || $uniVar := uniField && yField == "Y",... ) which, I think, isn't useful, just confusing. Would someone consider fixing this? -W