Case for the usefulness of ceteris paribus. The assumption is the economist's substitute for the controlled experiment that a laboratory scientist can perform. In the real economy, every variable moves at once — incomes, prices, expectations, policies — and isolating one causal relationship from observational data alone is extraordinarily difficult. By imposing ceteris paribus, the economist builds a theoretical model that states clearly: 'if only the price of this good changes, with incomes and other prices unchanged, the quantity demanded will fall'. This produces propositions that are clear, communicable and testable in principle, and lets a discipline build theory step by step. Without the assumption, economic theorising would have to describe every variable at once, which is intractable; with it, theory can proceed from simple two-variable relationships up to richer models in which assumptions are progressively relaxed.
Case against — the limitations. The usefulness has four significant limits. First, the assumption never holds exactly in reality: when a price changes, incomes, expectations and the prices of related goods are typically changing too, so the clean theoretical prediction is modified in practice. Second, important dynamic feedback effects are suppressed by the assumption — a change in the price of a good today changes expectations about future prices, which feeds back into current demand, and such interactions can be material. Third, over the long run and very long run, the variables held constant (technology, tastes, institutions) move substantially, so a ceteris paribus prediction can be a poor guide to actual outcomes over long horizons. Fourth, the assumption can be misused to insulate theories from refutation: if a prediction fails, the response 'the ceteris paribus condition did not hold' can protect a weak theory from honest challenge.
Reasoned judgement. The ceteris paribus assumption is indispensable at the theoretical stage of economic analysis — without it, propositions cannot be stated cleanly or tested logically, and the discipline could not have built up its stock of causal claims. But its usefulness is conditional on how it is applied: it is highly useful for short-run analysis of a single market where the suppressed variables are unlikely to move materially during the period in question, and it is far less reliable when applied to long-run questions or to policy problems in which the suppressed variables are themselves the subject of debate. The best practice is therefore to use ceteris paribus to build the theoretical proposition cleanly and then to relax the assumption systematically when applying the theory to a real-world question — asking which of the held-constant variables are plausibly moving at the same time and in which direction. On this reading the view 'ceteris paribus is useful' holds strongly for theory construction and short-run market analysis, holds moderately for intermediate applications where its limits are acknowledged, and holds weakly for long-run or system-wide questions where the suppressed interactions are the main story. The assumption is therefore a necessary tool but not a sufficient one — its usefulness depends on the care with which it is applied and the horizon of the question being asked.