Is BL Lacertae an orphan AGN? Multiband and spectroscopic constraints on the parent population

Aims. We have analysed optical spectra of BL Lacertae, the prototype of its blazar subclass, to verify the broad H alpha emission line detected more than a decade ago and its possible flux variation. We used the spectroscopic information to investigate the question of the BL Lacertae parent population. Methods. Low-and high-resolution optical spectra of BL Lacertae were acquired with the DOLORES spectrograph at the 3.58 m telescopio nazionale Galileo (TNG) during four nights in 2007-2008, when the source was in a relatively faint state. In three cases we were able to fit the complex Ha spectral range with multiple line components and to measure both the broad H alpha and several narrow emission line fluxes. Results. A critical comparison with previous results suggests that the broad H alpha flux has increased by about 50% in ten years. This might be due to an addition of gas in the broad line region (BLR), or to a strengthening of the disc luminosity, but such flux changes are not unusual in Broad Lined active nuclei. We estimated the BL Lacertae black hole mass by means of its relation with the bulge luminosity, finding 4-6 x 10(8) M(circle dot). The virial mass estimated from the spectroscopic data gives instead a value 20-30 times lower. An analysis of the disc and BLR properties in different AGNs suggests that this discrepancy is due to an underluminosity of the BL Lacertae BLR. Finally, we addressed the problem of the BL Lacertae parent population, comparing its isotropic quantities with those of other AGN classes. From the point of view of the narrow emission line spectrum, the source is located close to low-excitation radio galaxies. When one also considers its diffuse radio power, an association with FR I radio galaxies is severely questioned due to the lower radio luminosity (at a given line luminosity) of BL Lacertae. The narrow line and radio luminosities of BL Lacertae instead match those of a sample of miniature radio galaxies, which however do not show a BLR. Yet, if existing, misaligned BL Lacertae objects should have entered that sample. We also rule out the possibility that they were excluded because of a QSO optical appearance. Conclusions. The observational constraints suggest that BL Lacertae is caught in a short term transient stage, which does not leave a detectable evolutionary trace in the AGN population. We present a scenario that can account for the observed properties.