From: Fabian N. <fne...@we...> - 2005-09-08 18:56:23
|
> If the plasmocyte actively changes itself, it is "agent", otherwise > "patient" > I don't care that much about names, as long as the definitions are > clear. The names that I give below are the standard ones in semantic > role theories > What we need are additional roles for: > PATIENT: the role played such the entity playing that role is not > actively driving the process but undergoes some change > ENABLER: the role played such the entity playing that role is neither > AGENT, neither PATIENT but is required to participate or the process > would not happen > INSTRUMENT: the role played such the entity playing that role is used > by the AGENT to carry out the process, but the AGENT's action is not > directed towards it. > As I argued in my last email, it would be a mistake to introduce these distinctions in an ontology. They are mere linguistic distinctions. Consider the transcription of DNA. Is the DNA the agent which uses the RNA-polymerase as an instrument in order to produce mRNA? Or is the RNA-polymerase an agent and the DNA the patient of the process of building mRNA? This is a trick question. Although it is very common and often useful to speak about body parts as if they are acting, that's not literally true. Only higher organisms are able to act, because acting requires intention. Hence in the case of parts of cells there is nothing in reality which makes one participant an agent and the other a patient. To come up with another example, which shows that the criterions above don't work: If two water molecules arise from two hydrogen molecules and an oxygen molecule, who is the agent? Do the hydrogen molecules play a role that "is not actively driving the process but undergo some changes"? Are the oxygen molecules actively changing themselves? There are no good answers to the questions, because it is a bad question. 'Agent', 'Patient' and "Instrument" make sense as linguistic categories, but not as ontological categories. Best Fabian |