On the basis of literary arguments, Arkins (2011) "The meaning of Odi et amo in Catullus 85 came to the interesting conclusion that Odi et amo in the following famous poem by Catullus (LXXXV) cannot be simply translated as "I hate and I love" but rather as "I loathe her, I lust for her". His conclusion is indeed quite original and provocative if one considers the typical translations of this poem (e.g., for a selection of different translations, see Rose (2019) Odi et amo: A brief history in translation).
Odi et amo. Quare id faciam fortasse requiris.
Nescio sed fieri sentio et excrucior.
My specific linguistic question here is the following one: is it exceptional to find allegedly stative verbs like odisse and amare as reassumed by the agentive verbal phrase facere id? I'm asking this since in many languages it is often the case that stative verbs cannot be reassumed by the non-stative agentive verbal expression facere id 'to do it' (which is NOT the same as the grammaticalized expression 'to do so' in English). For example, the following Spanish and Catalan translations sound quite odd: Sp. “Odio y amo. ¿Por qué lo hago? (tal vez me preguntas…)” // Cat. “Odio i estimo. Per què ho faig? (tal vegada em preguntes…)”. Could you please let me know if the same happens in your native language? For example, TKR, in his comment below, also points out a similar semantic clash in Hebrew.
Here is some very sketchy linguistic background on so-called "psychological verbs". There is a typical tripartite classification of these verbs according to the syntactic function (and case, in Latin) of the experiencer argument: (I) nominative subject (timere-verbs), (II) accusative direct object (terrere-verbs) and (III) dative indirect object (placere-verbs). Typically, the first class is semantically analyzed as lacking an event, i.e., they are stative predicates. For example, see page 3 of Filip's (1996) work where you'll find the following statement: "Psychological predicates of the fear type are stative" (Filip 1996) NB: amare/'to love' and odisse/'to hate' are also assumed to belong to the timere/'to fear'-type/class, the verbs whose experiencer argument is the subject.
Following TKR's suggestion below, it would be interesting to find other similar contexts where psychological predicates of the 1st class above (verbs whose experiencer argument is the subject of the clause: e.g., amare, odisse, timere, etc.) are reassumed by facere id. Could you provide us with some attested examples of this linguistic phenomenon?
It is also VERY important to note that faciam and requiris are agentive predicates, compared to the verbs of the second verse, which are ALL non-agentive: nescio, fieri, sentio, and excrucior. So, in line with this distribution of agentive vs. non-agentive verbs in the 1st and 2nd verses, respectively, I was wondering if the use of faciam id to reassume odi and amo could be taken as a linguistic argument for extending the basic/lexical meaning of odisse and amare from a stative to a more agentive interpretation. If that were the case, this could also be taken as a non-subjective argument for Arkins's (2011) conclusion above in the sense that the translations of odi and amo in this poem should ideally incorporate a more agentive-like component, which is typically missing in 'I hate' and 'I love', respectively. Notice then that such an ingredient of agentivity would in turn nicely correspond more with the physical component argued for by Arkins (see the first link above) and less with the merely psychological lexical/basic meaning of these verbs.
It is also worth pointing out that a relevant chiasmatic structure has been said to be involved in this famous poem: cf. odi et amo -- excrucior; faciam -- (sentio) fieri; requiris -- nescio (alternatively, odi -- excrucior; amo -- sentio; faciam -- fieri; requiris -- nescio). In the light of the present question, note that this stylistic fact can also be put in correspondence with the linguistic/semantic contrast I have alluded to above: i.e. more agentive verbs (1st verse) -- non-agentive verbs (2nd verse).