The Pillars of Heracles is the name given by Plato to describe a maritime boundary marker of the ancient Greek world. According to his text, Atlantis lay just beyond or just before this boundary. However, strictly speaking, Plato does not call them ‘pillars’ but refers to them as stelai (pronounced “stee-lie”) and its singular Stele (pronounced “stee-lee”) which are the Greek words for stone slabs used as boundary or commemorative markers, not a reference to supportive columns. Rhys Carpenter favours the idea that the term when applied to the Strait of Gibraltar was used with the sense of boundary markers, indicating ”the limits of the Inner Sea that, for the Greeks, was the navigable world.”[221.156]
According to Aristotle, the Pillars of Heracles were also known by the earlier name of ‘Pillars of Briareus’. Plutarch places Briareus near Ogygia, from which we can assume that the Pillars of Heracles are close to Ogygia. Since Malta has been identified as Ogygia, it was not unreasonable, to conclude, as some do, that the Pillars were probably in the region of the Maltese Islands.
However, Ogygia has also been identified with one of the Faroe Islands in the North Atlantic by Felice Vinci[019.3], who then proposed that the Pillars of Heracles had also been located in that archipelago. John Larsen has made similar suggestions.
Furthermore, Aristotle also wrote that “outside the pillars of Heracles the sea is shallow owing to the mud, but calm, for it lies in a hollow.” This is not a description of the Atlantic that we know, which is not shallow, calm or lying in a hollow and which he refers to as a sea not an ocean.
Classical writers frequently refer to the Pillars without being in anyway specific regarding their location. Rosario Vieni has suggested that the Symplegades, at the Bosporus, encountered by Homer’s Argonauts were precursors of the Pillars of Heracles, although he settled on the Strait of Sicily as their location, before Sergio Frau adopted the same location.
However, there is little doubt that during the last centuries BC ‘the Pillars’ referred almost exclusively to the Strait of Gibraltar. The online encyclopaedia Wikipedia notes that Pillars were, in earlier times, identified with the Strait of Sicily, but from the time of Erastosthenes (c. 250 BC) the term was used to refer to the Strait of Gibraltar, reflecting the expansion of Greek maritime knowledge.
I consider it highly relevant that no writer prior to Eratosthenes had referred to the Pillars of Heracles being located at Gibraltar. It is not unreasonable to conclude that this silence reflects the lack of knowledge possessed by the ancient Greeks regarding the western Mediterranean, which only improved gradually, as their colonising and trading expanded westward.
Alessio Toscano has suggested that the Pillars were situated at the Strait of Otranto and that Plato’s ‘Atlantic’ was in fact the Adriatic Sea. A more distant location was proposed by Chechelnitsky who placed the ‘Pillars’ at the Bering Strait between the Chukchi and Seward peninsulas in Russia and the USA respectively.
It always seemed to me that when the Greeks began their expansion westward, they did so hugging the northern shores of the Mediterranean. Understandably, they would have taken the shortest route from the Greek mainland to the heel of Italy and later on to Sicily. As they progressed with their colonisation, new limits were set, and in time, exceeded. I suggest that these limits were each in turn designated the ‘Pillars of Heracles’ as they expanded further. I speculate that Capo Colonna (Cape of the Column) in Calabria may have been one of those boundaries. Interestingly, 18th century maps shown up to five islands near the cape that are no longer visible(g), suggesting the possibility that in ancient times they could have been even more extensive, creating a strait that might have matched Plato’s description. On the other hand, the Strait of Messina was one of the locations recorded as the site of the ‘Pillars’ and considering that mariners at that time preferred to stay close to the coast, I would opt for the Strait of Messina rather than the more frequently proposed Strait of Sicily.
An extensive collection of classical references to the ‘Pillars’ is to be found on the Internet(j).
However, the poet Pindar in the Third Nemean Ode would appear to have treated the Pillars as a metaphor for the limit of established Greek geographical knowledge (Olympian 3.43-45), a boundary that was never static. In 1778, Jean-Silvain Bailly was certain that the Pillars of Hercules were just “a name that denotes limits or boundaries.” [926.v2/293]
Dag Øistein Endsjø, a Norweigan professor, has added the use of the ‘Pillars’ as a metaphor to include the limits of human endeavour(d) and quotes the classicist, James S. Romm in support(e). My own view is that where the term may have initially referred to physical pillars, over time a metaphorical usage became the norm.
Paulino Zamarro has mapped 13 locations(f) identified as the ‘Pillars’ by classical authors (see map below) and expands on this further in his book. He identified Pori, a rocky islet north of the Greek island of Antikythera, as the location of the Pillars of Herakles.
The Pillars are assumed by some to refer to the Rock of Gibraltar in Europe and to Mt. Acha or Jebel Musain, which are near Ceuta in Morocco. Others prefer to accept them as a physical pair of pillars set up outside a temple.
The idea that geographical terms can radically change their location over time is illustrated by the name (H)esperia, which means ‘evening land’ or as we might say ‘land of the setting sun’, was originally used by Greeks to indicate Italy and later employed by Roman writers as a designation for Spain. It could be argued that the Greek use of this appellation could be an indication that when introduced they were not too sure what lay beyond Italy.
Herodotus (Hist. Bk II.44) refers to Heracles as a god of the Egyptians ‘from time immemorial’. He also visited a temple of Heracles in Tyre with two pillars, one of gold and the other emerald. According to the priests there it had stood for two thousand three hundred years or from approximately 2700 BC.
Heracles is clearly the Greek counterpart of the Phoenician god Melqart, who was the principal god of the Phoenician city of Tyre. Melqart was brought to the most successful Tyrian colony, Carthage and subsequently further west, where at least three temples dedicated to Melqart have been identified in ancient Spain, Gades, Ebusus, and Carthago Nova. Across the Strait in Morocco, the ancient Phoenician city of Lixus also has a temple to Melqart.
Gades (Cadiz) was originally named Gadir (walled city). It has been generally accepted that it was founded around 1100 BC, although hard evidence does not prove a date earlier than the 9th century BC. It is today regarded as the most ancient functioning city in Western Europe.
Pairs of free standing columns were apparently important in Phoenician temples and are also to be found in Egyptian temples as well as being part of Solomon’s temple (built by Phoenician craftsmen). Consequently the pillars of Melqart temple in Gades are considered by some to be the origin of the reference to the Pillars of Melqart and later of Heracles (by the Greeks) and Hercules (by the Romans) as applied to the Strait of Gibraltar.
*Spanuth dismisses those who have identified the red and white cliffs of Heligoland as the Pillars of Heracles, decrying the idea as a fallacy. He explains that “Natural rock formations were not what was originally meant by the Pillars of Heracles. Those at the Straits of Gibraltar were not, as one so often reads, the rocks to the north and south of the Straits, but two man-made pillars which stood before the temple of Heracles at Gades (present-day Cádiz) about 100 km north of the Straits.”*
The Pillars of Heracles usually play a critical part in the construction of any theory relating to the location of Atlantis. Even the authors of theories that have placed Plato’s island civilisation in such diverse locations as Antarctica, the North Sea or the South China Sea, have felt obliged to include an explanation for the ’Pillars’ within the framework of their particular hypothesis.
There is one location clue in Plato’s text (Tim.24e) that is often overlooked, namely, that the island of Atlantis was situated close to the Pillars of Heracles. Although it can be argued that Plato’s island was immediately before or after the Pillars, the text clearly implies proximity. This was pointed out by W.K.C. Guthrie in volume 5 of A History of Greek Philosophy [946.245] and independently endorsed by Joseph Warren Wells in The Book on Atlantis.
A number of alternative locations have been identified as being referred to in ancient times as the Pillars of Heracles. Robert Schoch writes “This distinctive name, taken from the most powerful hero of Greek mythology, was given to a number of ancient sites known in modern times by quite different appellations”. The Greeks, however, used the name Pillars of Heracles to mark other sites besides Gibraltar, some outside the Mediterranean – namely, the Canary Islands in the Atlantic and the Strait of Kerch dividing the Black Sea from the Sea of Azov – and even more inside – specifically, the Strait of Bonafaccio between Corsica and Sardinia, the Strait of Messina between mainland Italy and Sicily, the Greek Peleponnese, the mountainous coast of Tunisia, and the Nile Delta.
Arguably the most unusual suggestion this year has come fro Marco Goti in his book, The Island of Plato in which he identified the ‘Pillars’ in the Atlantic, being the basalt columns of the Giant’s Causeway in Northern Ireland in the west and their counterpart in Scotland’s Isle of Staffa in the east! However, this idea is not original, having been first mooted nearly seventy years ago by W.C. Beaumont(n).
Perhaps the first ‘modern’ writer to propose the eastern Mediterranean as the location for the ‘Pillars’ was Russian, Avraam Norov. He considered them to have been shrines drawing on both Greek and Arabic sources for his ideas.
G. Galanopoulos and E. Bacon suggest that the Pillars of Heracles were possibly associated with Melos, one of the Cyclades or Cape Maléa, the eastern promontory of the Gulf of Laconia. Both James Mavor and Rodney Castleden defend this view, which continues to have some support(p).?
Tacitus, the renowned Latin historian, in chapter 34 of Germania, clearly states that it was believed that the Pillars of Hercules were located near the Rhine in the territory of the Frisians.
Olof Rudbeck opted for a location further east in the Baltic at the Øresund strait between Sweden and Denmark.
Strabo, the Greek historian and geographer, records that Alexander the Great built an altar and ‘Pillars of Heracles’ at the eastern limit of his Empire. Pliny the Elder noted that in Sogdiana in modern Uzbekistan there was reputed to be an altar and ‘Pillars of Heracles’. Aristotle in de Mundo describes the north coast of Europe on the edge of a vast sea, beyond the Celts and the Scythians up to Sinus Gallicus and the Pillars of Heracles!
Ulrich Hofmann combines the Periplus of Pseudo-Scylax with the writings of Herodotus to build a credible argument for placing Atlantis in North Africa in Lake Tritonis, now occupied by the Chotts of modern Algeria and Tunisia. Consequently, Hoffmann places the Pillars at the Gulf of Gabés, which would put Malta to the east of them. He also argues that the Pillars were part of Atlantis rather than separate from it.
George Sarantitis presented a paper to the 2008 Atlantis Conference in which he also argued that the Pillars had been situated in the Gulf of Gabes[750.403]. He cites Strabo among others to highlight the multiplicity of locations that have been attributed to Pillars in ancient times.
Scylax of Caryanda describes(a) in his Periplus, a guide to the Mediterranean, the Maltese Islands as lying to the east of the Pillars of Heracles. The opinions of Hofmann and Sarantitis would certainly support this view.
Anton Mifsud argues that had the Pillars been located at Gibraltar the islands to the east would have been the Balearics. Mifsud also points out that the 1st century BC writer, Apollonius Rhodius, located the Strait of Heracles in ancient Syrtis Minor, now the Gulf of Gabés. Delisle de Sales placed the ‘Pillars’ not too far away at the Gulf of Tunis, the gateway to Carthage. Mifsud has now revised his opinion and in a December 2017 illustrated article(o) has identified the Maltese promontory of Ras ir-Raheb near Rabat, with its two enormous limestone columns as the Pillars of Herakles. This headland had originally been topped by a Temple of Herakles, confirmed by archaeologist, Professor Nicholas Vella.
Sergio Frau, in his recent book, published in Italy, insists that the Pillars were in fact located in the Strait of Sicily. He sees this location as according with the writings of Homer and Hesiod. He discusses in detail the reference by Herodotus to an island to the west of the Pillars, suggesting that the world ‘ocean’ had a different meaning than today and pointing out that elsewhere Herodotus refers to Sardinia as the largest island in the world. Following this lead Frau concluded that Atlantis was in fact located in Sardinia. Frau commented that Eratosthenes, circa 200 BC, was the first geographer to place the Pillars of Heracles at Gibraltar. He also quotes the earlier geographer Dicaearchus whose comments appear to also support a location near Malta. Antonio Usai , in his critique of Frau’s book has opted for the Pillars having been between the coast of Tunisia and the islands of Kerkennah.
It should be noted that many of the other known ‘Pillars of Heracles’ existed in the Eastern Mediterranean and Black Sea. This is possibly because until the middle of the 1st millennium BC the Greeks were, generally speaking, restricted to this region. It would appear that for the ancient Greeks, the Pillars of Heracles marked straits or promontories at the limits of their known world. These boundaries were extended further and further as their maritime capabilities improved. In the Late Bronze Age the Bosporus in the east and probably the Strait of Sicily in the west confined the Greeks. It was only shortly before Solon’s trip to Egypt that the Greek colony of Massilia (modern Marseilles) was founded and so, at last, the western limit of the Mediterranean was brought within the reach of Greek ships. Obviously as their range extended so too did the location of new Pillars and possibly led to the decline in the usage of the title at former boundaries, leaving us today with only the Strait of Gibraltar to carry the name.
Nevertheless, Thorwald C. Franke maintains that the westward shift of the ‘Pillars’ to Gibraltar occurred hundreds of years before Solon. He expanded on this at the 2008 Atlantis Conference and in his 2006 book on Herodotus.
Even Arthur C. Clarke suggested that there was evidence that the early Greeks did not originally refer to the Strait of Gibraltar as the Pillars of Heracles. Clarke did not cite his sources, but expressed a personal preference for the Strait of Messina.
Fundamentalist Atlantology, as proclaimed by the ‘prophet’ Donnelly in the 19th century, will accept no explanation other than that Plato was referring to ‘Pillars’ near Gibraltar. Certainly, it is perfectly clear that Plato MAY have been referring to the Strait of Gibraltar, but it is also clear that this was not the only location with that designation in ancient times. Consequently, if any of the alternatives mentioned above enable the construction of a new credible Atlantis location hypothesis, then it deserves careful rational consideration.
In 1913, Nicolae Densusianu proposed a location for the Pillars on the Danube, in ancient Dacia, modern Romania. Ranko Jakovljevic has recently expressed the view that the nearby Iron Gates section of the Danube in Serbia was the location of Atlantis. A paper presented to the 2008 Atlantis Conference by Ticleanu, Constantin & Nicolescu[750.375] has the ‘Pillars’ at the Iron Gates but place Atlantis a little further west on what is now the Pannonian Plain.
The late Arysio dos Santos claims claimed that “there was only one real pair of pillars: the ones that flank Sunda Strait in Indonesia”, in keeping with his Indonesian location for Atlantis. However, he does list (p.130) nine sites designated by ancient authorities as having been locations of ‘Pillars’, reinforcing the idea that the term was not exclusively applied to just one site.
Even more exotic locations such as Chott-el-Djerid in Tunisia, Bab-el-Mandeb(b) at the mouth of the Red Sea, the Strait of Hormuz(i) at the entrance to the Persian Gulf and even the Palk Strait between Sri Lanka and India have all been suggested at some stage as the ‘Pillars’.
George H. Cooper offered an even more outrageous solution when he wrote that Stonehenge in England were the original Pillars of Heracles.*In 2018, David L. Hildebrandt published Atlantis – The Awakening , in which he has endeavoured to do just that with a mass of material that he claims supports the idea of Atlantis in Britain and Stonehenge as the remnants of the Temple of Poseidon. He suggests that the the five trilithons represent the five sets of male twins, an idea voiced by Jürgen Spanuth and more recently by Dieter Braasch.*
Sometimes, in ancient Greek literature, this phrase refers to the strait between Sicily and the southern tip of Italy (a place which the Greeks did know well, having established colonies in Sicily and southern Italy). An indication of the level of confusion that existed in early geography and cartography is the fact that some ancient maps & texts mark the Mediterranean region west of the Strait of Sicily as “the Atlantic Ocean” and even state that Tyrrhenia is in the ‘Atlantic’!
Finally, my own conclusion regarding the location of the ‘Pillars’ referred to by Plato/Solon, is that a careful reading of Plato’s text shows clearly that they were located in the Central or Western Mediterranean. I base this view on (i) Critias 108 which states that the Atlantean war was between those that lived outside the Pillars of Heracles and those that lived within them and (ii) Critias 114 which declares that Atlantis held sway over the Western Mediterranean as far as Tyrrhenia in the north and up to the borders of Egypt in the south. Consequently, we can assume that west of Tyrrhenia and of Egypt were beyond the Pillars of Heracles. Depending on the exact location of the ancient borders of Tyrrhenia and Egypt, the Pillars could have been situated between Malta and Crete. This interpretation opens up the possibility of Malta, Sicily or Sardinia as prime candidates for the location of Atlantis.
Plato’s comments make little sense, if he was describing an attack by people outside the Pillars located at Gibraltar on those inside the Pillars, since at least half of the coastal territory, in both Europe and Africa, east of Gibraltar was already conquered. However, if the Pillars were located somewhere much further east, his comments make greater sense.
The Schoppes, in support of their Black Sea location for Atlantis, maintain that the Pillars were situated at the Bosporus and not Gibraltar. They contend “the maintained misinterpretation results from the fact that Herakles went to Iberia. At late Hellenistic and at Roman times Iberia was Spain. However, this leads to inconsistencies: After putting up the Pillars (supposed to be Gibraltar) Herakles put together a fleet to go to Iberia, he was still there!” The Schoppes point out that in the distant past ‘Iberia related to the land of an ethnic group to the east of the Black Sea.
Luana Monte, a supporter of the Minoan Hypothesis has also proposed a location at the mouth of the Nile Delta where the recently rediscovered sunken city of Herakleoin was situated. This identification appears to have been made in order to keep the Minoan Empire west of the ‘Pillars’.
Even more bizarre is the suggestion(c) that the ‘Pillars’ were in fact two bright stars in the western sky at the end of the last Age of Libra around 12,500 BC.
What is clear from all of the above is that the term Pillars of Heracles was, without doubt, applied to a variety of locations but Plato’s reference MIGHT relate to Gibraltar although equally strong if not stronger cases can be made for other sites at earlier dates. It is also plausible that at some point it also became a metaphor for any geographical limit.
Neville Chipulina, a Gibraltarian, has an interesting article(m) on the ‘history’ of the Pillars of Hercules and its association with Gibraltar.
Apart from any connection with Atlantis, it has been suggested that the vertical lines in the US dollar $ign represent the Pillars of Heracles!(l)