The reason you see this argument is because Latin has a complete case-structure, like many primitive languages, while English has no case structure. The case structure is appealing to people (this is the reason that it is introduced in the first place), since it allows sentences to match part of speech to a noun directly, not with little inconvenient word-attached tags. This case structure sounds superficially more "logical" than uncased languages like modern English, modern French, etc.
To approximate the effect, I will use the one remaining vestigial case in English--- the postmodifier "X-ward" to mean "to X". So you say
- I walk seaward.
- I walk to the sea.
It's a little archaic, but it's still obvious. Latin is chock-full of these cases, for all the prepositions you can imagine: "in X", "for X", "with X", etc, these all have little modifiers on the words to indicate that this is what X is doing in the sentence.
This is typical not just of Latin, but of most ancient or pre-written languages. The structure of these ancient languages then looks more complex then the structure of modern languages, leading to the degeneration paradox of linguistics: why are ancient languages more morphologically complex than modern ones?
In this linguistics stackexchange answer , I gave an answer to this question, one which I strongly believe, but which is not found in the literature as far as I know. It is mostly derived from my personal experience with Hebrew (which is formerly ancient, recently revived, and shed it's last two remaining cases, the locative and possessive, not too long ago to remember how it happened). The hypothesis I'm pushing is supported also by comparison of ancient epic poems (Gilgamesh vs. Odyssey), and from the example of Piraha, which is both non-recursive and insanely heavily cased.
Since Everett's work of 2005, it has become eminently clear that linguistic recursion is a much more recent invention than what was commonly supposed since at least 1957, and perhaps as far back as the renaissance. The Chomskian idea that all languages have a recursive grammar is dead wrong, Piraha lacks recursive grammar, full blown center recursion is absent from ancient Sanskrit, ancient Chinese (and a lot of modern Chinese), Australian languages like Warlpiri, ancient Hebrew, ancient Germanic languages, relatively recent Slavic languages (like midieval Polish), and more generally, any ancient language except ancient Greek and Latin, and some later Sanskrit, past Panini's time, which is approximately the 6th century BC. There are sporadic recursions in other ancient languages too, but the general rule is valid.
It seems that either the Greek or Sanskrit speakers invented recursion, and it spread horizontally from there, infecting language after language with recursion, until practically the entire old world was recursing. The recursion didn't touch Australia and America, although some native American languages had limited forms of recursion still, likely independently constructed.
Once you have recursion, complex case systems become unnatural, because in a recursive sentence, the noun objects can be very complicated to case. For example
- I walked ( to that rickety hard to maintain drug-den of a house which my father bought from the clan of fox-eating cave-people which settled on the shores of New Jersey in 1833).
Becomes, in some cased languages:
- I walked (that rickety hard to maintain drug-den of a houseward which my father bought from the clan of fox-eating cave people ...)
It's hard when the thing is embedded deep, and it requires a scan to move the noun-phrase to another position:
- That rickety hard to maintain drug-den of a house which my father bought from the clan of fox-eating cave-people which settled on the shores of New Jersey in 1833 is smelly.
That was easy. For the case example, I have to scan to find the "ward", and remove it. This is hard. So to accomodate easy recursion, languages replace cases with stand alone function words. You can date this to the time the language first becomes fully recursive among nearly all speakers, which usually coincides with the date of first universal literacy.
This is something I experienced personally with two case shedding events in modern Hebrew. The first concerns the locative ("to") case:
- Hu halach hahara.
In Hebrew means "He walked mountain-ward". It is never said this way in modern Hebrew, which replaces it with
- He halach lahar.
The "la" is attached to har, but that only means you can't insert adjectives before it, which you wouldn't do anyway because Hebrew puts adjectives after the noun.
The other case example, which is still in progress, is
- Shmi Ron.
This means "name-o-mine is Ron". This is not how people say it anymore (although it is still prescriptively encouraged, and bedevils Hebrew instruction). What you say is
- Hashem sheli Ron
"The name of me is Ron". I know that this transformation is easier to recurse with, since you often say:
- Hashem shel ha-ach hakatan sheli she nolad be-74 ze Gaby.
"The name of the brother the-little of-mine who was born in 74 it's Gaby" (my younger brother's name is Gaby). This easy recursion is obviously the reason the longer construction "hashem sheli" is preferred over "shmi", it is clear to any speaker that this is the driver of the case-shedding, and I inducted the principle from this experience.
Anyway, that's why Latin sounds more sophisticated. It has recursion, but it's tacked on and confined to upper class folks, it hasn't yet spread around to everyone, and once it does, it turns into uncased French, Italian, and Spanish.