I think the method identifier always must be qualified for this style (hence the name “infix notation”). The additional problem is that AFAIK in the REPL there’s nothing to qualify in with, i.e. there’s no identifier for the synthetic top level object, so you’d have to explicitly wrap the method declaration in a dedicated object in order to use this style.
scala> object Foo { def in(t:String) = print(t) }
object Foo
scala> Foo in "a"
a
However, use of infix syntax is discouraged anyway, except for specific cases (operators, HOF).