Quoting Josh Roth <[log in to unmask]>:

> Kar Marinam would treat the stone as an instrument in both
> instances, and a speaker would probably say that there is
> an agent too, that really caused the action, though we just
> don't know or want to say what it is (maybe a spirit or
> something).

Phaleran handles this in one of two ways.  First, it can use
the middle form of the transitive verb _rošwa_ "break", which
agrees with the patient:

  into       gellas     rošwaminti
  stone.INST window.ABS break.MID.3SgPfRe.S
  "The window broke (because of the stone)"

Alternatively, the instrument can be focused by a valence
changing operation _-ssa_, whereby the instrument receives
absolutive marking, and the patient is demoted to a dative
phrase that can be optionally deleted:

  î         (gellasuo) rošwassanti
  stone.ABS window.DAT break.INST.3SgPfRe.S
  "The stone broke (the window)."

In some dialects, this instrument-focuser has entirely replaced
the antipassive, which in the Standard normally takes the form
of the detransitivizer _-bro_ plus an antipassive prefix _s-_.

Thomas Wier
Dept. of Linguistics  "Nihil magis praestandum est quam ne pecorum ritu
University of Chicago sequamur antecedentium gregem, pergentes non qua
1010 E. 59th Street   eundum est, sed qua itur." -- Seneca
Chicago, IL 60637