LLNCS already defines a Proposition theorem environment, you should use \begin{proposition} ... \end{proposition}
instead of defining your own.
It has the advantage of properly localizing to various languages.
That said, the issue with prop
is not a conflict with LLNCS, but with thm-restate
. And it is not really that the environment prop
has been defined, but that there is a bug in how thmtools
is trying to latch onto the LLNCS macros and provide its own interpretation of stuff.
I've been able to isolate the bug to the file thm-llncs.sty
; the specific error is coming from the call to \thmt@newtheorem@predefinition
in the definition of \thmt@spnewtheoremiv
. (This part I am sure about.)
The problem, I think (the following is a guess, which may be incorrect), is that as part of the ...predefinition
, it calls ...autorefsetup
, which creates a counter-alias when sibling counters are used. But then the original \spnewtheorem
from the Springer LLNCS file is called, which tries to create the counter again.
One can check that without specifying a sibling counter, commands such as
\newtheorem{prop}{Proposition}
, or \declaretheorem[name=Proposition]{prop}
, or \spnewtheorem{prop}{Proposition}{\bfseries}{\itshape}
all work as expected, but as soon as you specify a sibling counter everything breaks.
You should file a bug with the thmtools
maintainers linking to this post and ask them to fix it.