This species is characterized by its thin-textured acuminate leaves, bilobed stipules, rather short, broadly pyramidal to corymbiform inflorescences, funnelform corollas with well developed projections on the abaxial (i.e., back) surfaces of the lobes, and didymous fruits. The identity of this species has long been confused. It was recognized for a while as Psychotria cornigera, but more often included within the circumscription of Psychotria bahiensis, sometimes as Psychotria bahiensis var. cornigera. However this species differs from Psychotria bahiensis in several characters. Palicourea didymocarpos has also been confused with Palicourea acuminata and Palicourea diminuta. The epithet "didymocarpos" is a noun in opposition and does not change its form to agree in gender with the genus name, but that is a common mis-spelling of it.
Palicourea didymocarpos as circumscribed here shows variation in several morphological features, but no clearly separable taxonomic groups are evident at present. Plants of the Ducke Reserve in Brazil that were identified as Psychotria cornigera have been reidentified as Palicourea rhodothamna, and do not represent Palicourea didymocarpos.
The name Ronabea didymocarpos was regarded by Taylor (2014) to be of unclear identity, because no corresponding specimen had been seen and the minimal description applies equally to Palicourea acuminata and Palicourea subcuspidata. The type specimen of Ronabea didymocarpos was located by Delprete & Kirkbride (2016), and corresponds to Palicourea subcuspidata as noted by them. Delprete & Kirkbirde overlooked the name Palicourea subcusptidata, but they cited its basionym Psychotria subcuspidata as new synonym of Palicourea didymocarpos (without indication it was a new placement) so this other combination is included in their taxonomy. Delprete & Kirkbride (2016: 418) also designated a nectotype for this name, "L.C.M. Richard s.n., P [P04017866], G-DC not traced". No type was cited in Candolle's work, which was based entirely on Richard's Rubiaceae treatment, at that time in press but unpublished. Richard cited the type of this name there as a collection of L.C. Richard, as Delprete & Kirkbride detailed. Delprete & Kirkbride regarded Candolle as the sole author of this name, and believe that only protologue information can be used to identify a type. Therefore they consider the type of Ronabea didymocarpos unknown . Other authors mostly (or all) regard Richard as the author of these names, which are then cited as "in DC." following Art. 46.2, and most also think that all available information should be used for typification. In this second case, the L.C. Richard specimen seems to correspond to the holotype of Ronabea didymocarpos.