Abstract. Nevanlinna's second main theorem is a far-reaching generalisation of Picard's Theorem concerning the value distribution of an arbitrary meromorphic function f . The theorem takes the form of an inequality containing a ramification term in which the zeros and poles of the derivative f ′ appear. In this paper we show that a similar result holds for special subfields of meromorphic functions where the derivative is replaced by a more general linear operator, such as higher-order differential operators and differential-difference operators. We subsequently derive generalisations of Picard's Theorem and the defect relations.
IntroductionNevanlinna theory studies the value distribution of meromorphic functions. Central to the classical theory is the second main theorem together with the related defect relations, which are powerful generalisations of Picard's Theorem. Nevanlinna's second main theorem uses the distribution of points in the closed disc |z| ≤ r at which a meromorphic function f takes certain prescribed values to bound the Nevanlinna characteristic T (r, f ). The theorem incorporates information from the distribution of zeros and poles of the derivative f ′ through a ramification term which ensures that when we count the preimages of the prescribed values we can ignore multiplicities.In this paper we derive an analogue of the second main theorem in which the derivative f → f ′ is replaced by an arbitrary linear operator f → L(f ) on any subfield N of the space of meromorphic functions such that m(r, L(f )/f ) = o(T (r, f )) for r in a large subset of (0, ∞). Examples of such operators are the derivative f (z) → f ′ (z), the shift f (z) → f (z + c) and the q-difference operator f (z) → f (qz) as well as combinations such asWe derive a generalisation of Picard's Theorem, which in essence says that if there are enough functions a j ∈ ker(L) that are small compared with f , then L(f ) is identically zero. Analogues of the defect relations are given and several examples are used to illustrate the strength of the results obtained.The work presented here extends earlier work on the shift [6] and q-difference [2] operators. Those works in turn grew out of a programme to use Nevanlinna theory as a tool for detecting and describing difference equations of "Painlevé type" [1,7]. The generalisations described in the present paper are motivated by preliminary studies of differential delay equations such as αf (z) + βf ′ (z) = f (z) [f (z + 1) − f (z − 1)] , (1.1)2000 Mathematics Subject Classification. Primary 30D35, Secondary 34M03, 34M05, 39A06.