The Ontology of Mythical Entities: Part 1
written by Scott Smith
The Ontology of Greek Mythical Figures, Take 1
This blog post is the first in a series of essays about the ontological status of Greek mythical figures and the difficulties in creating a stable database of Greek mythical entities. In this post, after offering some imperatives for creating a stable database of mythical names, I turn to some basic principles that inform our thinking in assigning “personhood”—and thus ontological status within a machine-readable world—to mythical figures.
As MANTO started to take on shape, it quickly became obvious that a stable database for mythical figures was necessary, one that could also serve as an authoritative resource for other digital projects based on Linked Open Data principles, as well as for traditional scholars working in the area of myth. An analgous tool is Pleiades, a crowd-sourced gazetteer of ancient places (Pleiades.stoa.org) that provides a stable URL for each place name, and therefore an unchanging location for information about it. Despite some lingering challenges, Pleiades remains an invaluable resource that offers a stable URI (Uniform Resource Identifier) for placenames. To take one example, Boeotian Orchomenos is assigned the Pleiades number 540987, while its Arcadian counterpart may be found at 570535. Even though individual gazetteers from which Pleiades draws its information may differ in the exact location of any one site (see the slightly different coordinates given by DARE and DARMC for Arcadian Orchomenos), the stability of a placename lies in that single URI on the Pleides website where the information is coordinated; this also allows new discoveries and refinements to be entered into the record without affecting the essential identity (or affecting other digital projects that are tapping into Pleiades, such as our own). Additionally, and this is extremely important for our own project, because the mere string of letters “Orchomenos” is insufficient by itself to denote what place is mentioned in a text, the text itself can be further encoded or marked up with specific URIs to distinguish between the Boeotian and Arcadian towns of the same name.
A stable and authoritative database for Greek mythical entities would be an equally beneficial asset for the digital classical community. Most importantly, a standardized URI for a mythical entity means that, instead of relying on a string of letters that make up a name, there is a stable ontological existence for the character or figure behind the name. With such an approach, characters with the same name can easily be differentiated and texts can be encoded to show the difference. For example, Ajax the Greater bears the MANTO ID: 8182067, while Ajax the Lesser can be found at MANTO ID: 8182068. Also, the several figures called “Lycos” would likewise carry a distinct ID (the value of this is indicated by the incredibly useful website Topostext, where the two Ajaxes are distinguished but “Lycus” remains aggregated).
Such a database would also allow one to easily systematize different spellings, different methods of transliteration, and multiple languages, e.g., Αἶας, Aias, Aeas, Aeax, Aiax, Ajax, Aiace. But there is also a great need for standardization; several lists of mythical figures exist, but there is little agreement on the arrangement. For instance, Lycos son of Pandion and brother of Aigeus is #8 in Brill’s New Pauly, #6 in Grimal’s Dictionary of Classical Mythology and Roscher’s Lexicon, #21 in RE, #2 in LIMC, and is #5 in Dizionario Etymologico della Mitologia Greca and Lycus7 in the online Greek Mythology Link.
And yet, unlike real “biological” people, mythological figures are not necessarily ontologically singular. As scholars of myth are intimately aware, mythical narratives, genealogies, locations and even names are vigorously contested and prone to manipulation and variation. Furthermore, even though many mythical figures take on biological features and can be subjected to something approximating a biography (Achilles, for example), many names that populate the mythical world cannot be said to refer to a “character” at all or be said to have agency in the mythical storyworld. Thus, organizing Greek mythical figures as a simple prosopography is bound to end up in failure.
Let us start with a simple question: do mythical figures exist? By this I do not mean whether Odysseus or Heracles was a “real” character in that there was a flesh-and-blood person underlying the vast number of stories that accrued around the names. It is obvious that, even if there was a “real” Heracles, this would not account for the variety of accounts that feature very different versions of “Herakles” in them. Instead, I mean: is it productive to speak of “Heracles” as an ontologically identifiable entity even if non-existent in a biological way? Similarly, can one say that an obviously unreal figure like the Chimaira exists ontologically?
Recently, the question of whether fictional (e.g. Sherlock Holmes), mythical (e.g. the planet Vulcan), and purely imaginary figures exist has been the subject of vigorous inquiry by philosophers. Based on Alexius Meinong’s theory of objects, which posits that existence is not a property of an object, an emerging consensus posits, generally, that characters and other non-existent entities are abstract objects that are by-products of the process of creating narratives, plots, and novels, or the result of mistaken or imaginary thinking. That there exists what might be called “conceptual” entities is not new, but has ancient precedent. If Seneca is a reasonable guide (Ep. 58.15), some Stoics had already anticipated a super-category quid that included real (corporeal) and unreal (incorporeal) objects. Among the latter were entities, such as Giants and Centaurs, that were created by mental “false reasoning” (quae animo succurunt ... quicquid ... falsa cogitatione formatum habere aliquam imaginem coepit, quamvis non habeat substantiam). We may extrapolate as follows: because Giants and Centaurs are written about and depicted on art, and because those names refer to stable properties that are generally agreed upon (for example, Centaurs have a horse body but human torso and head), they belong to an ontological category of things that exist but are not corporeal.
To put it another way, it is quite different to say that a name that refers to a stable set of mental images does not exist on the one hand, and a random string of letters does not exist on the other.
1. Centaurs do not exist 2. Hgkskdtifwr does not exist
Even if one claims that Centaurs do not exist, how can the following sentences have clear differences in truth value?
3. Centaurs have the body of a horse and a human torso and head
4. Centaurs have the body of a chicken and the head of a monkey
It is obvious that the first sentence is true in a way that the second is not. In short, as Maria Reicher (2013: 159) has plainly stated, “there is no satisfactory way to avoid an ontological commitment to fictitious characters.” Conversely, there are good reasons to accept such a status: these entities can be discussed, referred to, drawn, depicted, and are even protected under law. US Copyright law has determined that fictional characters can be protected separately from their underlying works if the characters are sufficiently unique, distinctive, and to merit such protection. To take an example, Wonder Woman is copyright protected and is owned by DC Comics. In order to claim such protection, however, there must be a clear set of stable characteristics that determine what is being protected (name, costume, etc.). It is, however, not necessary for there to be an underlying living figure for there to a meaningful reference. To say that Wonder Woman is less of an entity than George Washington, Napoleon, or Hugh Jackman neglects the fact that representations of even real people do not have to accurately reflect the existing underlying object. In other words, persons can be said to exist outside of their lived lives just as characters exist outside of their narrative structures.
It is that stable set of characteristics which is the focus of our mythical ontology and on which I will focus below. As for the Centaurs, they do exhibit some variability in presentation. In early depictions, for instance, they have human front legs, but by the classical period they get four horse legs, and the stories about their origins and genealogies can differ. In the Roman period female Centaurs appear. And yet, these variations do not overpower the basic recognizability of the species. In other words, they have stable and recognizable data points that allow us to talk about them productively as an ontological category.
When it comes to how the Greek mythical storyworld operates, however, it is important to understand that not all mythical entities are created equal or have the same function in the storyworld. To give a brief statement here (to which we will return to treat in more detail later): entities with developed characters such as Perseus and Achilles are clearly analogous to biological persons in a way that an eponymic figure like Plataia is not. In this blog post I will present two case studies of the former (biographical) type to drive home a basic principle that will resonate in later posts: differences in genealogical details or narrative acts are not enough in and of themselves to create separate entities.
Case Study #1: The Many Fathers of Io
Consider the assertions found in a Zitatennest (“Nest of References”) from Apollodorus’ Library about Io (2.5):
1. Io is the daughter of Iasos (no citation; cf. Paus. 2.16.1)
2. Io is the daughter of Inachus (Castor, tragedians; cf. Paus. 1.25.1, 3.18.13)
3. Io is the daughter of Peiren (Acusilaus, Hesiod)
These alternate genealogies do not affect Io’s role in the narrative story. She remains a priestess of Hera, becomes the object of Zeus’ affections, is turned into a cow, roams over the world, ends up in Egypt, and gives birth to Epaphus. In other words, the only contested data is her genealogical connections to other parts of the mythological tradition; the rest of her identity is assured by the stable data shared in multiple versions. A similar example may be seen in Daphne, whose father is variously given in our ancient sources. Are there multiple figures named Io or Daphne? Surely not, but in the index to Rocha-Pereira’s edition of Pausanias, Io is given two entries based on variant parentage: one as daugther of Iasos, and one as daughter of Inachos. Here, the primacy of place given to parentage is here taken to the extreme, but it does show how often scholars of myth tend to privilege parentage (admittedly, they follow the lead of ancient mythographers). Take a set of entries from Pierre Grimal’s Penguin Dictonary of Classical Mythology, all of which start with the parents:
Euchenor: “A son of the soothsayer Polyidus (1)...
Eudorus: “A son of Hermes and Polymela...”
Eumaeus: “The son of Ctesius...”
Eumelus: “The son of Admetus and Alcestis...”
Eumolpus: “According to the most common tradition, the son of Poseidon and Chione...”
Case Study #2: Two Figures Named Icarios
In the above case study we considered stable data points (Io as priestess, raped by Zeus, wanders to Egypt) and unstable data points (paternity). When considering the aspects of myth that can be said to be stable or unstable, we may point to three main factors: 1) genealogy; 2) geography; and 3) the role in the narrative storyworld. To emphasize these, let us consider briefly the two mythical figures named Icarios. We find one Icarios whose data points are all stable: the Athenian Icarios is always father of Erigone and receives wine from Dionysos (MANTO ID: 8188569). When we turn to Laconian Icarios, we find a great deal more variation in terms of genealogy and family connections, as well as in terms of his actions in the storyworld. We outline those differences below, with the stable data points highlighted in bold:
Laconian Icarios
1) Genealogical Ties
son of Oibalus or Perieres, husband of
either Polycaste
with whom he is father of Alyzeus, Leucadios and Penelope
or Periboia
with whom he is father of five children + Penelope
2) Geographical Origin
Spartan
3) Narrative Actions
either helps Hippocoon expel Tyndareus
or was himself exiled with Tyndareus by Hippocoon
The stable data points, then, consist solely of his geographical association with Sparta, and that he is exclusively the father of Penelope. Yet, these stable data points—along with consistent even if contested relationships with the same figures (Hippocoon, Tyndareus)—are enough to characterize Laconian Icarios as a separate and stable ontological entity (MANTO ID: 8188570). In each of these cases (Io and the Icarioi) we are dealing with what might be called a “biological entity,” whose actions impact the broader storyworld. In the next post, we will consider the ontological status of non-agent mythical figures with a particular focus on eponymic figures.