Does Dirac-Born-Infeld modification of quadratic theories really matter?

We study the consequences of further modification of f(R,RμνRμν,RμνσρRμνσρ)/f(R) theories by means of the Dirac-Born-Infeld procedure, which is the replacement of f by λ(√(1+2f/λ)-1) (the free parameter λ fixes an additional energy scale). We pay special attention to the definition of masses of the linearized propagating degrees of freedom because they are important to judge the stability of the linearization around vacuum background spaces. In this context we discuss the subtleties associated with expanding f(R,RμνRμν,RμνσρRμνσρ) Lagrangians around maximally symmetric spaces of constant curvature, as well as with equivalence of the linearized Lagrangian to a scalar-tensor theory. Investigation of the consequences of applying the Dirac-Born-Infeld (DBI) strategy to further modify quadratic theories on the stability of de Sitter vacuum, as well as its impact on the cosmological dynamics, are the main concern of this paper. We show that (i) although the DBI deformation does not affect the Ostrogradski stability, other important instabilities such as the Ricci and scalar-tachyon ones, may be indeed surmounted (sometimes at the cost of renouncing to the original motivation of the DBI strategy, to avoid singularities), and (ii) DBI transforming the original theory broadens its possibilities to do cosmology since the asymptotic structure of the DBI-dual theory is richer than in the standard case. In particular, either the dimension of the phase space is increased, or there appear bifurcations in the control-parameter space.