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, an indivisible blob that marks the verb as past‐tense‐first‐person‐singular, the Turkish equivalent is ‑d‑im, where the ‐d‐ marks the tense and the ‑im independently carries the subject agreement. If you switch to the past‐tense‐first‐person‐plural, Italian has a completely unrelated blob, ‑ammo, while Turkish just modifies the subject‐agreement component to give ‑d‑imiz. It was recognised well before Zamenhof came along that the Turkish approach makes a better groundplan for the morphology of a constructed international auxiliary language, since it avoids the need to memorise combinatorial tables of arbitrary grammatical endings. In Zamenhof's neighbourhood this groundplan was represented by the non‐Indo‐European languages 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 he didn't recognise the distinction between
derivational and inflectional morphology.
It's the derivational system that allows Esperanto morphemes to
link together in long German‐style chains like the compound noun
= “tea‐pot‐cosy collectors' club‐house”. Inflectional
morphemes don't get much beyond
= “exiles” – and even there, the ‑it
is a composite tense/aspect/
The third major option (which has influenced Esperanto's verbal system there) is the “analytic” groundplan, which consists of avoiding inflectional morphology, like subject‐agreement affixes, in favour of separate words, like explicit subject pronouns. (When it eliminates derivational morphology as well, that's the fully “isolating” groundplan.) It turns out that a case can easily be made for thoroughly analytic 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 verb modifiers depending on your mother‐tongue habits, and there's also nothing to stop you treating them as if they were attached to the verb as inflections.
- Inflecting grammars are basically alien to speakers of languages with no inflecting features; but the barrier is one‐way, because there are no languages entirely lacking in analytic features.
- Analytic languages are by no means obscure – for a start they dominate Eastern Asia from China through to Indonesia, and English is itself largely analytic; 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 together as a community of native‐speakers); they are “designed” by the innate preference babies have for a complete but easily learnable grammar, and they tend overwhelmingly to use analytic rather than heavily inflecting 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 nalunaarasuartaatilioqatigiiffissualiulersaaleraluallaraminngooq = “it seems that they were well into the process of talking about founding an association for the establishment of a telegraph station”… but this is rarely proposed as a model for an auxiliary language!)