ISA meets Lara: an incremental word space model for cognitively plausible simulations of semantic learning