ON THE MAIN INVARIANT OF ELEMENTS ALGEBRAIC OVER A HENSELIAN VALUED FIELD (Q4784801)

In this paper the authors characterize the Henselian valued fields \((K,v)\) for which to each \(\alpha \in \overline{K}\setminus K\), there corresponds \(\beta \in \overline{K}\) satisfying \([K(\beta):K]<[K(\alpha):K]\) and \(\delta _ K(\alpha)=\overline{v}(\alpha-\beta)\). Here \(\overline{K}\) is an algebraic closure of \(K\), \(\overline{v}\) is the unique extension of \(v\) to \(\overline{K}\) and \(\delta _ K(\alpha) = \sup\{\overline{v}(\alpha-\beta)\mid \beta \in \overline{K}, [K(\beta):K]<[K(\alpha): K]\}\) considered as an element of the Dedekind order completion of the value group of \(K\). They prove that these fields are exactly those such that for each \(\theta\in \overline{K}\), \(K(\theta)/K\) is a defectless extension with respect to the valuation obtained by restricting \(\overline{v}\) (a finite extension \((L,w)\) of a Henselian valued field \((K,v)\) is said to be defectless if \([L:K]= ef\) where \(e\) and \(f\) are the index of ramification and the residual degree of \(w/v\) respectively). This result gives rise to the question: For a Henselian field \((K,v)\), if \(K(\theta)/K\) is defectless for each \(\theta \in \overline{K}\), then is it true that every finite extension of \((K,v)\) is defectless? By means of an example in the last section of the paper, the authors show that in general the answer to this question is no.