Faced with the charge that the design of Esperanto's grammar is parochial, the nearest thing Zamenhof's apologists have to counter‐evidence is the fact that Esperanto's morphology was avowedly influenced by “agglutinative” languages such as Turkish rather than the “fusional” model dominant in Europe. What this means is that where Italian verbs have endings such as ‐ai, which signals past‐tense‐first‐person‐singular in one indivisible blob, the Turkish equivalent is ‐d‑im, where the ‐d‐ marks the tense and the ‐im carries the person agreement. It was recognised well before Zamenhof came along that this makes a better groundplan for the morphology of a constructed international auxiliary language, since it avoids the need to memorise combinatorial tables of grammatical endings. In Zamenhof's neighbourhood this groundplan was represented by Finnish, Estonian, and Hungarian – all predominantly agglutinative, though they look very europeanised when put alongside other examples such as Korean, Luganda, or Quechua.
Zamenhof eagerly adopted the concept of discrete invariable building‐blocks; but there are two kinds of block – “derivational”, used to build new vocabulary items, and “inflectional”, used to signpost grammatical features. It's mostly the derivational system that links Esperanto roots together in long but regular chains, as in the compound noun te‑krucho‑mufo‐kolekt‑ist‐ar‑ej‑o “tea‐pot‐cosy collectors' club‐house”. The inflectional ones never get much beyond ekzil‐it‑o‑j‑n “exiles” (and even there, the ‐it is a sort of crypto‐fusional composite tense/voice marker, while the ‐ojn is modelled on the fusional declension patterns of classical Greek). So in effect Esperanto is like a version of German with its affixes de‐fused, not like a paradigmatically agglutinative language. A whole‐heartedly Turkish‐style constructed language would handle all the various modal, reflexive, conditional, or aspectual forms of verbs by stacking verb‐endings, so that (for instance) “I won't have been seen”, mi ne estos vidita, would instead use perfective, passive, future, negative, and person‐agreement suffixes to form something like, say, vid‐iv‐at‐ur‑en‑im.
The third major option (which has influenced Esperanto's verbal system there) is the “isolating” groundplan, which consists of eliminating affixes (“re‐straight‐en‐ed”) in favour of multiword phrases (“did make straight again”). It turns out that a case can easily be made for thoroughly isolating solutions being more convenient for more people:
- They have better extensibility – that is, they can be taken from sparse simplicity to rich subtlety just by learning more vocabulary.
- They accommodate different agreement‐marking instincts well – you can include or omit or relocate (e.g.) aspect‐marking words depending on your mother‐tongue habits (and there's nothing to stop you using “did‐make‐straight‐again” as if it was a single agglutinated word).
- Affixing grammars are basically alien to speakers of languages with no affixing features; but the barrier is one‐way, because there are no languages entirely lacking in isolating features.
- Isolating languages are by no means obscure – they dominate East Asia from China through to Indonesia, and English is itself largely isolating; that already adds up to something like forty percent of the human race (and climbing)!
The natural equivalent of the artificial auxiliary language is the “creole” (which is what a sub‐linguistic pidgin turns into once children start growing up as native‐speakers); they are “designed” by the innate preference babies have for a complete but easily learnable grammar, and they tend overwhelmingly to use isolating rather than agglutinating groundplans.
(And for the sake of completeness I should also mention the fourth basic groundplan: “polysynthetic” grammars are exemplified by the West Greenlandic one‐word sentence ininnukalaarniarlungaana, “the thing is, I'm going to my room for a bit”… but this is rarely proposed as a model for an auxiliary language!)