## Categorial compositionality continued: A category theory explanation for quasi-systematicity

### BibTeX

@MISC{Phillips_categorialcompositionality,

author = {Steven Phillips and William H. Wilson},

title = {Categorial compositionality continued: A category theory explanation for quasi-systematicity},

year = {}

}

### OpenURL

### Abstract

universal construction The classical account for systematicity of human cognition supposes: (1) syntactically compositional representations; and (2) processes that are sensitive to their structure. The problem with this account is that there is no explanation as to why these two components must be compatible, other than by ad hoc assumption (convention) to exclude nonsystematic variants that, e.g., mix prefix and postfix concatenative compositional schemes. Recently, we proposed an alternative explanation (Phillips & Wilson, 2010) without ad hoc assumptions, using a branch of mathematics, called category theory. In this paper, we extend our explanation to domains that are quasi-systematic (e.g., language), where the domain includes some but not all possible combinations of constituents. The central category-theoretic construct is an adjunction involving pullbacks, where the focus is on the relations between processes, rather than the representations. In so far as cognition is systematic, the basic building blocks of cognitive architecture are adjunctions by our theory.

### Citations

745 |
Learning representations by back-propagating errors
- Rumelhart, Hinton, et al.
- 1986
(Show Context)
Citation Context ...of systematicity specifically, see Aizawa (2003). At this point, modellers may think to augment their theory with some sort of learning principle, such as is commonly incorporated into connectionist (=-=Rumelhart, Hinton, & Williams, 1986-=-), and Bayesian modeling (Tenenbaum, Griffiths, & Kemp, 2006). However, connectionist and Bayesian approaches suffer the same shortcoming: while both are capable of configuring architectures with the ... |

573 | Connectionism and Cognitive Architecture: A Critical Analysis’, Cognition
- Fodor, Pylyshyn
- 1988
(Show Context)
Citation Context ...of cognitive capacities. This property of cognitive architecture (i.e., the collection of basic processes and modes of composition that together generate cognitive behaviour) is called systematicity (=-=Fodor & Pylyshyn, 1988-=-), and the problem posed for a theory of cognition is to explain why systematicity is a necessary consequence of the assumptions and principles embodied by the architecture that the proposed theory po... |

172 | Basic Category Theory for Computer Scientists - Pierce - 1991 |

106 | Categories for the Working Mathematician, 2nd edition - Lane - 1998 |

77 | H.: Conceptual Mathematics: A First Introduction to Categories - Lawvere, Schanuel - 1997 |

33 | Category theory - Awodey - 2006 |

32 | Realization is Universal - Goguen - 1973 |

25 | LOT 2: The Language of Thought Revisited
- Fodor
- 2008
(Show Context)
Citation Context ... providing the key advance concerning the foundations of cognitive science, overcoming the problems with 352associativism by suggesting that cognitive processes are instead (syntactic) computations (=-=Fodor, 2008-=-). Turing’s (classical) solution works well for computational systems, because the correspondence between the processes for constructing compositional representations and the processes for accessing t... |

8 |
The Systematicity Arguments
- Aizawa
- 2003
(Show Context)
Citation Context ... theory of cognition is to explain why systematicity is a necessary consequence of the assumptions and principles embodied by the architecture that the proposed theory posits (Fodor & Pylyshyn, 1988; =-=Aizawa, 2003-=-). The classical explanation derives from the principle of classical compositionality, which says that cognitive representations and processes are constructed from a combinatorial syntax and semantics... |

7 |
2004): “On the Systematicity of Language and Thought
- Johnson
(Show Context)
Citation Context ...ften incorporate different types of constraints, including syntactic, phonetic, semantic, and pragmatic constraints that may further restrict the group of capacities that are intrinsically connected (=-=Johnson, 2004-=-). For example, English-speakers say John put his gear down, but not John stowed his gear down, even though they say John put his gear away, or John stowed his gear away (see Johnson, 2004). In this s... |

6 | Category theory as a conceptual tool in the study of cognition - Magnan, Reyes - 1994 |

2 | Applying category theory to improve the performance of a neural architecture - Healy, Olinger, et al. - 2009 |

2 |
Categorical Compositionality: A Category Theory Explanation for the Systematicity
- PHILLIPS, WILSON
- 2010
(Show Context)
Citation Context ...ther than by ad hoc assumption (convention) to exclude nonsystematic variants that, e.g., mix prefix and postfix concatenative compositional schemes. Recently, we proposed an alternative explanation (=-=Phillips & Wilson, 2010-=-) without ad hoc assumptions, using a branch of mathematics, called category theory. In this paper, we extend our explanation to domains that are quasi-systematic (e.g., language), where the domain in... |

2 | What do Transitive Inference and Class Inclusion have in common? Categorical (co)products and cognitive development - Phillips, Wilson, et al. - 2009 |

1 | A new foundation for representation in cognitive and brain science: Category theory and the hippocampus. Unpublished doctoral dissertation, Escuela Tecnica Superior de Ingenierus Industriales - Gomez - 2010 |