Realistic Social Networks in the Táin

Enkidu

Interesting report on a study here:
http://www.irishtimes.com/newspaper/breaking/2012/0726/breaking6.html

Apparently the relationships between the greater than 400 characters in the Táin is realistic once you factor out six characters (such as Medb). An interesting result is that Beowulf is very realistic, followed by the Ilyiad. The Táin is not realistic until you remove six superhero characters (which have an effect on the social network similar to superheroes from Marvel).

However Shakespeare's Richard III is less socially realistic than all of them, only slightly better than Lord of the Rings.

Original paper here:
http://iopscience.iop.org/0295-5075/99/2/28002

slowburner

Enkidu wrote: »

Interesting report on a study here:
http://www.irishtimes.com/newspaper/breaking/2012/0726/breaking6.html

Apparently the relationships between the greater than 400 characters in the Táin is realistic once you factor out six characters (such as Medb). An interesting result is that Beowulf is very realistic, followed by the Ilyiad. The Táin is not realistic until you remove six superhero characters (which have an effect on the social network similar to superheroes from Marvel).

However Shakespeare's Richard III is less socially realistic than all of them, only slightly better than Lord of the Rings.

Original paper here:
http://iopscience.iop.org/0295-5075/99/2/28002

Makes you wonder what the original tales were based on (if they truly have become distorted in the telling).

The notion that mathematical (statistical?) analysis could be used to separate fact from fiction in the grey areas of history is a very exciting concept indeed.

Copied this to the archaeology forum too - hope you don't mind.

Nitochris

Enkidu wrote: »

Interesting report on a study here:
http://www.irishtimes.com/newspaper/breaking/2012/0726/breaking6.html

Apparently the relationships between the greater than 400 characters in the Táin is realistic once you factor out six characters (such as Medb). An interesting result is that Beowulf is very realistic, followed by the Ilyiad. The Táin is not realistic until you remove six superhero characters (which have an effect on the social network similar to superheroes from Marvel).

However Shakespeare's Richard III is less socially realistic than all of them, only slightly better than Lord of the Rings.

Original paper here:
http://iopscience.iop.org/0295-5075/99/2/28002

I'm relying on the newspaper but it reminds me of K.H. Jackson's Window on the Iron Age in which he claimed an oral antiquity for the Táin Bó Cúailnge going back to around the fourth century. Cúchulainn and otherwise are also assumed to be mythical, but on the strength of the similarities between Roman record of Gaulish culture and the world featured in the text itself it provides, or so argues Jackson, a window on the Iron age.

The above is the nativist claim and has been disputed by revisionists, among others J.P. Mallory who in The World of Cúchulainn: The Archaeology of the Táin Bó Cúailngeshows that the material culture and non-Irish fauna appearing in the eco-system reflects the present for the composition of the Táin Bó Cúailnge, that is the early medieval era. He also points out that at the same time the writers of the Ulster Cycle attempted to reconstruct the past, avoiding what they knew to be anachronisms.

Nitochris

slowburner wrote: »

Makes you wonder what the original tales were based on (if they truly have become distorted in the telling).

Garrett Olmsted (1992) reckons it features on the Gundestrup Cauldron, he has also translated Conailla Medb Míchura despite being an earlier version of The Táin a certain leading figure is absent, the late Dáithí Ó hÓgáin suggested on the basis of this poem that Fiacc son of Fergus was the original hero.

pmacc

Pretty cool to see my work being discussed here! Also Enkidu you might be interested to know that the Gilgamesh is high on my priorities to analyse.

Nitochris wrote: »

Garrett Olmsted (1992) reckons it features on the Gundestrup Cauldron, he has also translated Conailla Medb Míchura despite being an earlier version of The Táin a certain leading figure is absent, the late Dáithí Ó hÓgáin suggested on the basis of this poem that Fiacc son of Fergus was the original hero.

Thanks for this, I have not heard that Fiacc son of Fergus might have been the original hero, it would be interesting to compare the different networks to see if the societies are similar from a statistical point of view.

Enkidu

pmacc wrote: »

Pretty cool to see my work being discussed here!

I assume this is Pádraig MacCarron then. Very kind of you to take the time to post here.

Also Enkidu you might be interested to know that the Gilgamesh is high on my priorities to analyse.

Very interested!:)

pmacc

Enkidu wrote: »

I assume this is Pádraig MacCarron then. Very kind of you to take the time to post here.

That's right yeah, been a long time since I posted here!

If anyone's interested I presented this work at an archaeology conference a few months ago so I explain most of the concepts in detail as it's not geared towards a physics audience. You can view it here.

slowburner

Fascinating 'pan-disciplinary' material which should provoke thought and discussion in quite a few other forums; such as here, here, or here (already sneaked a copy over to archaeology).
I found the assortativity concept particularly intriguing.

Enkidu

pmacc wrote: »

That's right yeah, been a long time since I posted here!

If anyone's interested I presented this work at an archaeology conference a few months ago so I explain most of the concepts in detail as it's not geared towards a physics audience. You can view it here.

Thanks for the link to that very interesting talk. The concepts are explained very well.

Technical aside:
I was wondering, with regards to the power law distribution. Is there something different about social networks with a exponent less than one?
(The continuous version would give 1/k, whose integral doesn't converge, of course you are dealing with discrete probabilities with some ultimate upper bound on their range, so this doesn't really matter, but still just curious.)

pmacc

slowburner wrote: »

Fascinating 'pan-disciplinary' material which should provoke thought and discussion in quite a few other forums; such as here, here, or here (already sneaked a copy over to archaeology).
I found the assortativity concept particularly intriguing.

Assortativity's nice alright, here we only used it to test degree-degree assortativity (degree = how many links you have) but you can test for assortativity for any property of a node, for example if you had the ages of each person you could see if people tend to associate with people of a similar age (which, in general, I'm sure they do)

Enkidu wrote: »

Thanks for the link to that very interesting talk. The concepts are explained very well.

Technical aside:
I was wondering, with regards to the power law distribution. Is there something different about social networks with a exponent less than one?
(The continuous version would give 1/k, whose integral doesn't converge, of course you are dealing with discrete probabilities with some ultimate upper bound on their range, so this doesn't really matter, but still just curious.)

For a pure power-law probability distribution of the form

p(k) ~ k^-gamma

the exponent gamma is defined to be a positive constant greater than 1, this is so the probability can't be greater than one (most of the time you'll do cumulative distributions anyway so when you calculate the exponent you've to add one to the value to get the real exponent if that makes sense).

If the exponent is less than 2 it has infinite mean, between 2 and 3 it has a mean but no variance (scale free), when I talk about it not converging in the presentation I mean the variance and not the probability distribution, I didn't make that clear actually I just realised but at the same time I wanted to keep it straight forward and not talk about things like first and second moments! Also when I say infinite here I just mean very large, as you say we're dealing with discrete rather than continuous values here. Usually networks with a high exponent tend to be quite robust (transport networks for example, they don't have heavy tails in their distributions).

In some social networks there will be a cut-off however so it won't be a pure power law, it will be a truncated one, in this case the exponent can take any value (within reason). It'll look like

p(k) = k^(-gamma) exp(-k/kappa)

where kappa is some cut-off for degree where the exponential will entirely dominate (I think). These are quite common but the exponent from the power-law part still tends to be in a similar range to non-truncated ones.

dubhthach

Of course with regards to the "super-hero" characters such as Medb (Meadhbh / Méabh) there is an element of been probably deities "humanised" into a narrative. After all there was the goddess "Medb Lethderg" at Tara who is associated with the kingship. Rath Maeve been one of the sites that is just south of main Tara complex.

Cú Chulainn likewise has an aspect of deity about him, given the réamhscéal (pretale) about his conception/birth -- triple conception, the god Lugh been his actual father etc. Who were the other members of the "superhero 6" (I see Conchobhar Mac Nessa is one)?

pmacc

dubhthach wrote: »

Of course with regards to the "super-hero" characters such as Medb (Meadhbh / Méabh) there is an element of been probably deities "humanised" into a narrative. After all there was the goddess "Medb Lethderg" at Tara who is associated with the kingship. Rath Maeve been one of the sites that is just south of main Tara complex.

Cú Chulainn likewise has an aspect of deity about him, given the réamhscéal (pretale) about his conception/birth -- triple conception, the god Lugh been his actual father etc. Who were the other members of the "superhero 6" (I see Conchobhar Mac Nessa is one)?

5 of them are fairly obvious (going to go with Kinsella's spelling here): Cúchulainn, Conchobor, Fergus, Medb and Ailill. The sixth isn't quite so clear, this is Finnchad Fer Benn, one of Conchobor's sons, at one point Conchobor tells him to summon all the warriors of Ulster and names about 100 characters (about half of whom we've met before), making the links here is a little dodgy, we assumed that Finnchad must have known all these warriors and they knew him, arguably this is not the case however removing Finnchad alone makes very little difference to the network so we left it in for completeness.

You might be interested to know that in terms of degree, Dubhthach was tenth highest.

dubhthach

pmacc wrote: »

5 of them are fairly obvious (going to go with Kinsella's spelling here): Cúchulainn, Conchobor, Fergus, Medb and Ailill. The sixth isn't quite so clear, this is Finnchad Fer Benn, one of Conchobor's sons, at one point Conchobor tells him to summon all the warriors of Ulster and names about 100 characters (about half of whom we've met before), making the links here is a little dodgy, we assumed that Finnchad must have known all these warriors and they knew him, arguably this is not the case however removing Finnchad alone makes very little difference to the network so we left it in for completeness.

You might be interested to know that in terms of degree, Dubhthach was tenth highest.

I'm not surprised obviously it's nice having a character in the national epic who shares your "ancestral name" -- Dubthach Dóeltenga -- Dubhthach of the "beetle tongue" (literally Black Tongue). Modern Irish Dóel = Daol, tenga = Teanga

Obviously in old irish Dubh is written as Dub. It's kinda a doubly dark name as Dubhthach implies darkness in it's own right.

One of the obvious anarchorism in the Táin is the use of the name Connacht / Connachta. The province after all been named after the Connachta whose name implies descent from Conn of the Hundred battles, they along with the Uí Néill form the Dál Cuinn.

The pseudo-histories place Conn as living in 2nd century AD which if we go with the syncretic histories is after the time that Táin supposed happened.

Before the arrival of the Connachta west of the Shannon you supposedly had the Fir Ol nEchmacht, thence the old name of province been Cóiced Ol nEchmacht (Cóiced = Cúige in M.Irish eg. fifth/province)

The mention of Conailla Medb Míchura (Maeve makes evil contracts) is interesting, as it centers the action in Tara (not Cruachán in Connacht) that and it probably dates from the mid 7th century.

Obviously Kinsella's translation is based off what's called Recension 1 (Old Irish) pieced together from the fragmentary parts in Leabhar na hUidhre and the Book of Lecan. From what I've recalled reading elsewhere it's reckoned that this oldest version is compiled from at least two source versions from older now lost manuscripts. Linguistically sections of it are supposedly datable to the 8th century.

This ties in with general era that the "syncretic historians" were basically rewriting Irish history to tie it into wider Christian world etc.