Imagine a conlang built on the principles of a philosophical language
consisting of monosyllabic words with broad basic meanings. However,
in getting from the protolang to the "finished" conlang you would use
processes of fusion, erosion, vowel mutation, etc., that would
eventually completely disguise the philosophical nature of the roots.
The initial roots would be monosyllabic names of concrete objects,
actions, and parts of objects like body parts. The lexicon would be
completely concrete with no words for abstract concepts like
"happiness", or "friendship", and no adjectives or adverbs, let alone
prepositions or articles. Verb tenses would, of course, be out of the
question. There might not even be any pronouns to begin with, and
counting words might be initially limited to "one, two, three, many".
The lexicon should probably contain pointing words "this" (near me),
"that" (near you), "yonder" (far from both of us), also monosyllabic.
Sentences would consist of agent/subject before object/patient with
action either between those two or after both, with no initial
preference for either. Two or more consecutive sentences that share
the same agent/subject could be chained together with a comma
replacing the repeated agent/subject. That way "Man walk. Man leave
hut. Man come river." could be condensed into "Man walk, leave hut,
come river." That would be the only sentence template initially
Words may only be coined by putting together philosophical roots, but
other processes can then be applied to the result. Words may also be
put together in set phrases which can then become idioms which might
then be shortened or simplified to create new words. Words can change
roles so that "Man walk, leave hut, come river." could come, after
many centuries, to mean "Man walk from hut to river." as the original
monosyllable "leave" ends up being used as if it were a preposition
"from". Perhaps with another new coinage to replace the verb "leave"
with the compound "go-leave" where the word for "leave" has come to
mean "from". Likewise, "come" eventually takes on the meaning "to", so
that "come river" eventually means "to river".
Abstract nouns like "friendship" and "beauty" would initially be
created with compound words or set phrases used metaphorically.
Adjectives would come from concrete nouns used as exemplars of the
trait described. "leaf" or "grass" might be used to mean "green", and
"elephant" might be used for "large". Later, those words my experience
vowel shifts when used in an adjectival sense in order to distinguish
them from the concrete objects, but their roots in the concrete might
still be detectable. In any case it should be possible to document the
history of a word back to its original protolang monosyllables, even
if the final word no longer bears any resemblance to its original
For example, suppose the pointing word "ho" means "this/here", and the
word "yi" means "person". We could coin a first person singular
pronoun from the set phrase "ho yi" = "this person" Later that might
be eroded to "hoi", and later still to "oi". Only the vowels remain
from the original "ho yi", but the conlang now has a first person
singular pronoun with a documented pedigree, rather than some
arbitrary word coined out of thin air.
Similarly, "you" might begin as "ta yi" = "that (near you) person",
and "he/she" might be "pen yi" for "yonder person". Later, "ta yi"
might, over many centuries, become "tai" -> "tau" -> "dau" as long as
each step is something that can be justified in some reasonable
manner, and is not just some radical, arbitrary change. "this", "that"