In Geography IV 3:3, following Asclepiades the Myrlean, Strabo describes a temple of Athena located in a city called Ulyssea, where shields and ship beaks had been nailed in memory of Ulysses. Though Strabo meantions it when describing the Mediterranean coast of the Iberian Peninsula, the settlement was interpreted by some as being Olisipo. A case in point is Damião de Góis, a 16th century Portuguese humanist who published a description of Lisbon in 1554, including the legend of the city’s founding by Ulysses and its temple to Minerva, just like in the description by Asclepiades (Urbis Olisiponis Descriptio, II 6). No word, however, on traces of it in the 1500s and that’s true for the 21st century as well.
The existence of Olisipo’s temple to the Goddess of One Thousand Crafts is more assumed than archaeologically known. There are news of the discovery of a statue of Minerva on the castle hill (Moita 1994, 52), but it’s a very vague reference and I was unable to track down the exact document that mentions the finding. And then there’s the phenomenon of religious recycling, whereby the same sacred site is used by different religions in different periods, which is well-known in other places across Europe, but again only assumed in this case. As it stands, in the absence of written sources and archaeological excavations to confirm it, it is believed that there was a temple to Minerva on the castle hill, more precisely on the site of today’s church of the Holy Cross. The modern building is an 18th century construction, but it replaced an older church that was destroyed in the 1755 earthquake and which dated back to the Portuguese conquest of the city in 1147. Before that, it would have been a mosque, a possibility recently substantiated by the uncovering of remains of a rich Arab neighbourhood, likely for the governing elite and bureaucrats of Muslim Lisbon. And thus we reach the theory of the location of Minerva’s temple, which starts with a medieval church that would have been a mosque and makes a further leap into the past to assume a Roman stage in the history of that religious site. It’s not impossible, but it’s not certain either; and the doubt will remain until excavations either prove or debunk the theory.
Jupiter too is assumed by some to have had a temple on the castle hill, though no archaeological traces of it have been found. The theory follows similar steps as Minerva’s and takes the old royal chapel, which would have been dedicated to Saint Michael and Saint Barbara (who is associated with thunderstorms), to assume that it might have replaced a Roman cult site (Moita 1994, 52). There are written references to a votive inscription to Jupiter Optimus Maximus, but the 16th or 17th century document that mentions it fails to add exact information on where it was found apart from having been in Lisbon (Vieira da Silva 1944, 246; n. 127). Other than that, there’s nothing else to support a location on the castle hill.
There is, however, another votive inscription to the King of the Gods and it was found in 1842 in the monastery of Outer Saint Vincent, named so because it was outside the city walls when it was originally built in the 12th century. It reads: IOVI PRO. SALVTEM M. CASSI. F. … RMI M. IVLIVS PRIMVS V.S.L.A. (Vieira da Silva 1944, 192-3; n. 81). There are news of more archaeological traces in the same place, but dedicated to imperial figures: when the monastery was fully refashioned in the 16th century, an inscription to Emperor Vespasian was said to have surfaced from the rubble in 1582 (Vieira da Silva 1944, n. 80) and another one from the city to the daughter of Emperor Trajan is said to have been inside the church around the same time (1944, 65; 82). There are also news of an inscription that mentions a flaminica in the section of the 14th century wall next to Outer Saint Vincent (1944; n. 83). All of this combined leaves one wondering about the existence of an imperial cult site on that hill top, with which Jupiter might have been associated.
One question remains, though: could it have been located outside the Roman walls? Olisipo is known to have been fortified during the Roman period, at first by Brutus when he took the settlement in 138 BCE and again later also in the 4th century, but the exact location of both walls throughout the urban area is unknown. A few traces have been found here in there, like under the house of spikes (red dot on the map), where remains of the Moorish wall overlap vestiges of a semi-circular Roman bastion (Guardado da Silva 2008, 46). Assumingly, the 10th century Arab wall (green line) used only parts of the old defence line, thereby leaving out several Roman buildings (outlined in white) which will be mentioned further on.