Even though one may easily stress our ignorance of how the central engine in AGN operates, some observational and theoretical progress has probably already been made. Reprocessing seems to be significant. perhaps giving rise to the low-ionization lines and partly to the IR, optical-UV and X-ray continua. The X-ray reflection features provide evidence of cold, optically thick matter, and are consistent with a flattened geometry, such as an accretion disk. It remains to be seen how the BLR, the NLR and the intermediate torus and possibly the jets are related to the central accretion disk. The customary Shakura-Sunyaev approach seems inadequate in this exploration process, since several effects are not included. However, the more relevant slim disks will probably also require modification, since the soft X-ray excess and the hard X-ray power law seem difficult to reproduce. One solution could be to include a hot, optically thin plasma in or near the inner disk. A significant part of the liberated gravitational energy may then be dissipated in this region, thereby producing the hard X-rays which subsequently get reprocessed into lower-energy radiation in several steps. Both the geometry of and the dominant physical processes in the hot phase are unsettled, but may involve Comptonization, pair production or magnetic fields in a corona or disk annulus.
As a first step towards physical understanding of AGN, the existence of both accretion disks and black holes needs to be settled. If a method of estimating the central mass is to be reliable, alternatives which involve the central engine may be more appropriate. Due to the inherent problems with UV-continuum fits, essentially two possibilities remain: interpretation of X-ray continuum variability or of Fe fluorescence line profile variations. Since the latter alternative awaits adequate spectral resolution, the best current evidence of black holes in AGN seems to come from variability studies, as exemplified by NGC 6814.
The striking similarities between spectral features (blue bump, X-ray reflection features, hard X-ray power law, quasi-periodic variability and so on) observed in AGN and X-ray binaries may indicate the need for a general model, applicable to both stellar and extra-galactic systems. If combined with the results from forthcoming X-ray satellites (such as XMM and AXAF), such a model should be able to constrain or even deduce physical conditions in compact sources.
Prof. M. Abramowicz, Prof. B. Pagel and Prof. R. Svensson are thanked for their comments on the manuscript. Financial support from the Royal Swedish Academy of Sciences, the Royal Physiographical Society of Lund and the Swedish Natural Science Research Council is gratefully acknowledged.
DK-2100 Copenhagen Ø, Denmark