On defining irreducibility (Q4489997)

The paper shows that irreducibility is not a definable property for algebraic varieties over the real field \({\mathbb{R}}\). In detail, for every \(p \geq 2\), there is no first order sentence \(F_p\) in the language \(\{+, -, \cdot, 0, 1 \}\) enlarged by a 1-ary relation symbol \(I_p\) such that, for any algebraic variety \(V\) of \({\mathbb{R}}^p\), if one interprets \(I_p\) by membership to \(V\), then \(F_p\) is true in \({\mathbb{R}}\) if and only if \(V\) is irreducible. In fact the author proves that defining irreducibility contradicts the o-minimality of \({\mathbb{R}}^{\text{exp}}\) (the expansion of the reals by the exponential function). Accordingly, irreducibility remains undefinable even if we replace \({\mathbb{R}}\) by \({\mathbb{R}}^{\text{exp}}\) above. Furthermore, the same proof still works if one deals with arbitrary real closed fields (instead of \({\mathbb{R}}\)) or one allows parameters. A similar argument shows that, for \(p \geq 2\), algebraic varieties over the reals cannot be defined among semialgebraic sets. Parallel results for algebraically closed fields have been proved by the author and some coworkers (Chapuis, Hrushovski, Poizat) in other recent papers.