Grammar of a language (that is phonology or morphosyntax) does not depend on on geography/climate and mostly also on culture/religion. You can see that easily in Europe where you have all sorts of languages within more or less comparable area (highly synthetic Ancient Greek, Latin and Czech, analytic Romance languages or German, agglutinative languages like Basque or Turkish, isolating languages like English).
Furthermore most of these languages have the same origin, i.e. most of them are Indo-European, and current reconstructions indicate this was a highly inflected language, yet we have very wide variety of language types originating from it.
Geographically, the only influence that you can find is the one of language contact but you ruled it out by postulating "isolated geographical setting".
Culture/religion can have influence on language structure but typically on higher levels than grammar, i.e. semantics, pragmatics and vocabulary (and vice versa - the language has an influence on the culture). Generally speaking, we can say culture represents the way we see the world, therefore it includes the mental constructs we use to interpret it, which are typically reflected in a language somehow.
The influences in grammar or usually much less frequent but if there is some super important cultural feature, it might be somehow reflected. E.g. in Japan, there used to be (and still partially is) very important element of politeness and social stratification. This is reflected in Japanese language, where you have many honorific pronouns (compared to typical European that has typically 2 levels in singular: 2nd person singular pronoun for relations of acquaintance and more, 2nd person plural pronoun for politeness; and no distinction in plural) and even some honorific verbs, maybe there are even some different morphological structures.
Another interesting example is spatial and temporal metaphors - in European languages, the timeflow metaphor is based on you moving through the world, i.e. you move forward and therefore what was (the past) is behind you, whereas what is going to be (the future) is in front of you. In certain Native American languages (Quechua, Aymara,...), it is the time that moves around you, not you in it, and it moves the opposite direction - you can see/remember what was (the past), therefore it lies in front of you, whereas what is going to come (the future) remains to be seen and therefore it lies behind you.
I will not go as far as to say this example is purely accidental since there might be some hidden factors making more likely for a culture to develop the metaphor one way or the other but you can certainly look at it as accidental.
However, what I do believe that has an indirect influence, is the... let's say, state of civilisation. If you have a large and interconnected civilisation with lots of people who can write (say the Roman Empire), it is likely that the development of the linguistic/phonological structures is delayed significantly (even for centuries) because it is much more difficult for a change to prevail. When the civilisation breaks down into smaller, isolated units and the writing is limited to the elites, the evolution kicks off and the languages mutate wildly and become mutually uncomprehensible (say spoken Romance languages before Latin relexification)
Regarding the literature, I would start with something from Sapir & Whorf, the linguists that are considered the fathers of linguistic typology. However I do not know any serious sources that would link linguistic structures to particular extra-linguistic phenomena (there were some attempts in Soviet linguistics but they were as short-lived as their authors and very bad science). Concerning the evolution, this is a rather new outlook so there is not that much, but you can actually look into evolutionary biology where some linguists like to look for inspiration (e.g. Flegr: Frozen Evolution gives the maths for development of a genetic feature based on the size of the population).
