Choose what you are measuring

Complexity can refer to the size of a sound inventory, the number of inflectional distinctions, the irregularity of a paradigm, syntactic patterns or processing demands. These are different properties. A language with few case forms can still have intricate constructions, while a language with many forms can organize them predictably. A single informal adjective such as simple conceals which dimension is being compared.

Do not assume a perfect balance

It is tempting to claim that every difference in one domain must be exactly compensated elsewhere. That is an empirical hypothesis, not a principle established by counting a few features. WALS discusses this problem when comparing sound inventories and related dimensions. Human languages support rich communication without needing identical numerical complexity on every proposed measure. Structural differences should be described directly rather than forced into a ranking.

Difficulty has a learner

How difficult a language feels depends partly on what a learner already knows, which modality and writing system are involved, and what level of use is being learned. That experience is not a universal property of the language or its speakers. The laboratory compares consequences of particular rules: how many forms are distinguished, which ambiguities remain, and what a construction requires. It does not award a best-language score.

THE THREAD TO KEEP

Complexity needs a defined measure; learning difficulty needs a learner and a task.

Try it for yourself

Compare two teaching systems on one stated dimension, then list what the comparison leaves unmeasured.