This seemed weird to be because "talnaa" (to fry) is transitive.
I think
talnaa is one of the rare-ish Hindustani verbs that can be both transitive and intransitive. To me at least, it sounds fine to say things like
pakauRe tal chuke haiN, or
pakauRe jab tal jaaeN... Also, the intransitive listing does show up in
Platts.
(In any case, this is sort of besides the point: as has already been pointed out by others, it's fine and not at all weird to use the perfect participle of transitive verbs attributively in front of nouns.)
* ublaa and ubaalaa are a "non-causative / causative" verb pair, rather than a "transitive/intransitive" pair. Non-causative/causative and transitive/intransitive are overlapping, but not the same concepts. One would say they are "orthogonal" concepts.
I'm not really sure how you're using "causative" here...? In particular, if transitivity and causativity were really orthogonal concepts, there would be a thing that is a "intransitive causative," and I don't know what that would mean.
That said, a convenient (and probably the typical) way of thinking about Hindustani verbs is that they tend to come in groups of three-ish, each of the three having a different
valency (ie, number of nominal arguments the verb needs to form a complete thought). The typical situation is a group of three that consists of a verb of valency 1 (ie, an intransitive), another of valency 2 (ie, a transitive), and then a third of valency 3 (and this one is typically what's called the "causative," in my experience?). The group
ubalnaa/ubaalnaa/ubalvaanaa fits this typical paradigm well:
panii ubalaa = the water boiled
A ne paanii ubaalaa = A boiled water
A ne B se paanii ubalvaayaa = A had B boil water
But this typical paradigm has lots of exceptions. For instance, one could probably regard
khaanaa/khilaanaa/khilvaanaa as a group, but here the valencies are 2, 3, and 4, respectively. (This 4-valent "double causative" usage of
khilvaanaa is somewhat rare, so it could be that acceptability judgements vary...?)
A ne kuchh khaayaa = A ate something
A ne B ko kuchh khilaayaa = A fed B something
A ne B se C ko kuchh khilvaayaa = A had B feed C something
Another kind of exception is exhibited by the group
likhnaa/likhaanaa/likhvaanaa, where to me it sounds like the latter two both have valency 3 and are more or less synonymous:
A ne kuchh likhaa = A wrote something
A ne B se kuchh likhaayaa/likhvaayaa = A had B write something
* my understanding is that the "passive voice", per se, doesn't exist in Hindustani. A general idea of "passiveness" or experiencing a change of state is conveyed by adding "jaanaa" to other verbs, but essentially whether or not a verb can be read in a passive sense or not, is given by its dictionary definition.
I'm not sure it's fair to say that the passive voice doesn't exist in Hindustani. It's probably worth making sure that we're drawing a distinction between verb stem +
jaanaa (which is a compound verb construction, and which is not what seems to be under discussion here), and then the perfective participle of transitive verb +
jaanaa construction. It is this latter construction that sounds decidedly passive to me. The fact that most transitive Hindustani verbs already have intransitive counterparts means that there's often little reason to use the passive construction (ie, perfective participle +
jaanaa), since it usually sounds far more natural to just use the intransitive counterpart. But in situations where there is no intransitive counterpart, one can and would use the passive construction:
kaam kiyaa gayaa (for "the work was done," since
karnaa has no intransitive counterpart),
khat likhaa gayaa (for "the letter was written," since
likhnaa has no intransitive counterpart), etc.
* I am not 100% sure that Hindustani speakers automatically perceive a past participle as "passive", the same way that western-Indoeuropean language speakers do.
I guess this is the kind of thing that could vary from person to person, but I would tend agree with this assertion. The typical way of forming the past tense uses the perfective participle means, so I don't really have a strong association between the perfective participle and the passive. To reiterate my previous point though, I do perceive the perfective participle of transitive verb +
jaanaa construction as passive.
