I'd say that Interlingua is "scientific" in the sense that it was
constructed on basis of scientific research. Look at it this way: imagine
that sometime in late 1949 IALA had decided to junk the idea of an IAL, and
simply publish their research. Essentially, they'd have produced a work of
philology: a scholarly survey of neo-Latin and pan-European Wanderwoerter.

 Now, suppose the heading for each entry was "standardised" form of the
different variants - would that be science, linguistic engineering, or
pseudo-science? Well, to say that you can model a standard form according to
objective methodology is neither scientific not unscientific, although it is
compatible with scientific activity.

However it is legitimate philology to note that LA "causa" has mutated into
two separate words (in terms of form/meaning) in the modern european
languages, which can be modelled as CAUSA and COSA. (Interestingly enough,
De "Sache" has shadowed the Latin development, from 'cause/legal case' to
'thing' - illustrating the convergence that underlies the idea of SAE.) It's
also philology to note that in many cases the modern-day forms have
completely lost some original phonetic features so that the "theoretical
ancestor" may not be the actual etymological form. That's something the
prototyping methodology could show you.

The Interlingua-English Dictionary, as published, is not a scientific work,
but it was based on scientific work, IMO. I don't make the same claims to
Ia's grammar, however. It seems fairly impressionistic, although Gode & Co.
tried to follow, in spirit, the same approach they'd taken to vocabulary.