Subjects/Objects in chemical situations

[mereolog]

Concepts in the vicinity of substance constituency use the subject/object metaphors, e.g., 'has object role', 'has subject role' to describe the situation of the mereological nature.

For instance, AmlodipineBesylateConstituency hasObjectRole AmlodipineAsActiveMoiety. The definition of hasObjectRole then implies that AmlodipineAsActiveMoiety identifies the thing that is affected by or is secondary argument wrt to AmlodipineBesylateConstituency. At the same time, AmlodipineBesylateConstituency hasObjectRole hasSubjectRole AmlodipineBesylateAsSubject, which means that AmlodipineBesylateAsSubject identifies the thing that is being discussed or is the main topic in a specific role in AmlodipineBesylateConstituency.

These metaphors may be seen as adequate in some specific domains that depend on human intentionality (like finance), but they sound odd in when applied to the chemical reality.

In general, I do not think that situations have any topics or components that are being discussed or described. Secondly, AmlodipineBesylateConstituency like any other situation does not seem to have any primary (affecting) or secondary (affected) components. Considering the examples of situations in IDMP and FIBO one may perhaps say that one thing "in" a situation affects another thing "from" this situation, but these relations are somewhat symmetric. So even if one may say that a situation have subjects and objects, then its "subjects" affect its "objects" to the same degree in which its "objects" affect its "subjects".

I would suggest making this explicit and speaking about constituents and compounds or simply about parts and wholes. Then AmlodipineAsActiveMoiety and AmlodipineBesylateAsSubject would have, respectively, the part/constituent and the whole/compound role in AmlodipineBesylateConstituency.

Perhaps a native speaker may find better terms than "constituents/compounds" or "parts/wholes", but irrespective of this detail I would suggest replacing the current language-inspired language (subjects, objects, topics, etc., by the mereology-driven language (parts, wholes, etc.), at least in the case of substances and their components.

[ElisaKendall]

@mereolog We are still working through this - I agree that the names of the properties aren't perfect in some cases, and have refined a number of the properties linking to "object roles" so far. But, not all of the things that need to be attached to a given constituency are role-related - there are contextual details as well. If we didn't need the context information we might have been able to avoid reifying the concepts in something like a "situation" altogether. And, it's possible that what we really need is a new construct whereby there is no need for a role on the 'subject' side of the reification, but it connects to possibly multiple roles on the object side. There are also some constraints that need more refinement, such as for the active ingredient relation that has a separate reference substance as its basis of strength, the relationship with the reference substance has to be included, but in other active ingredient role cases it is not required. These may need to be bundled so that we can represent complex drugs that have multiple active ingredients, whose active ingredient roles may differ.

In other words, "stay tuned" - I'm working on adding a couple more examples that are more complicated, which I hope to have done by next week, so that we can see if we can simplify on the "subject" side and invent a new reification thing, rather than a situation, that supports more complexity in terms of context and "object side" content.