Ancient accounts for etymologies can be enlightening and at the same time be quite misleading. What's still true today: in all cases that involve folk etymology, the real, linguistic etymology may be completely unrelated.
Here we are dealing with myths as well. Just looking ta the myths surrounding the origin and founding of Rome itself, which is much better documented than anything about the Marsi. But if we are to trust the fontes it begs the question, which myth to believe?
For the Marsi we have knowledge of then existing in the time of the early and middle republic, which is a problematic field of historiography as such. We also have knowledge about them from writers that are quite removed on the timeline from the people in question.
- Tiberius Catius Asconius Silius Italicus (c. 28 – c. 103 CE)
Gaius Julius Solinus, Latin grammarian and compiler, probably flourished in the early 3rd century AD. Historical scholar Theodor Mommsen dates him to the middle of the 3rd century.
Pliny the Elder (/ˈplɪni/; born Gaius Plinius Secundus, AD 23–79)
Then we have to look at the ethnonym and Latin exonym of Marsi for a people of Oscan-Umbrian heritage which came into more heavy contact with the Latin Romans in the 4th century BC.
For the historical development of langauge contact between the two grous: Rex Ervin Wallace: "The Sabellian Languages", Dissertation, The Ohio State University, 1984. (PDF)
Now consider all these circumstances: early contact, generally mythological explanation expected, "Marsi, known as tenacious fighters, closely allied with Rome, after Samnite wars", late folk etymological explanations of exonym found in primary sources.
This led to the following conclusion:
Several words should be said about the ethnonym of Marsians. Marsians, ancient inhabitants of Italy, associated their ethnonym with Lidian king Marsyas.2
In Bayet‟s opinion, the legend about Marsyas‟ coming to Italy, which is included in the Roman Caco myth, is an absolutely extraneous element and was created under Greek and Etruscan influence in Capua in approximately the IV century B.C.3
The other parallel version explaining Marsians eponym is more trustworthy, and links Marsians with Circe‟s (Medea‟s) son, Marsos.4
Some scholars consider archaism of this version quite convincing since Circe‟s or Medea‟s participation (here I mean the goddesses of fecundity in general, whose names were later substituted for Circe) in formation of ethnonyms of different Italian tribes is rather widespread.5
2 Plin., nat., III 108, Silv., VIII, 502-504, Sol., II, 6, Osid., Orig., IX, 2, 88.
3 J. Bayet, Les Origines de l‟Hercule romain, Paris, 1926, 214.
4 Plin., nat., XXV, 11; Gell, XVI, 1.
5 Cesare Letta, I Marsi e il fucino nell' antichità, Centro studi e documentazione sull' Italia,
Romana. Monografie supplemente degli Atti 3, Milano, 1972, 53.
Src: Ekaterine Kobakhidze: "The Issue of Descent of the Deity Mars", Phasis 8, 2005. (PDF)
It is thus not incorrect on Wikipedia to paraphrase
Tribe in the Central Italian Appenninus near Lacus Fucinus (Str. 5,2,1; Ptol. 3,1,57), regarded as brave and warlike (Str. 5,4,2; Plin. HN 3,106; Liv. 8,29,4; Verg. G. 2,167; cf. the etymological derivation from ‘Mars’). As descendants of a son of Circe they allegedly were immune to snake poison (Plin. HN 7,15; 21,78; 25,11; 28,30; Gell. NA 16,11,1; cf. Cic. Div. 1,132; 2,70; Hor. Epod. 17,29; Hor. Carm. 2,20,18). Characteristic of the M. was the cultivation of vegetables and herbs on their land (cf. Plin. HN 25,48; Columella 2,9,8; 6,5,3; 12,10,1). The area of the M. was bisected by the via Valeria (Str. 5,3,11). The following towns of the M. are known: Anxa (modern Gallipoli), Antinum, Lucus Fucens (modern Luco), Marruvium (San Benedetto; Plin. HN 3,106), and Archippe (Plin. HN 3,108). Early on the M. entered into friendly relations with Rome: In 308/7 BC the consuls supported them against the Samnites (Diod. Sic. 20,44,8); in 304/3 BC a treaty with Rome was concluded (Diod. Sic. 20,101,5; Liv. 9,45,18). The Social War 3 (91-89 BC) was called bellum Marsicum after its instigators (see Diod. Sic. 37,1; Str. 5,4,2; Diod. Sic. 37,2; Liv. per. 72-76; Vell. Pat. 2,15).
Nissen 2, 454
R. Sclocchi, Storia dei M. 1-3, 1911
G. Devoto, Gli antichi Italici, 1931, 335ff.
C. Letta, I M. e il Fucino nell'antichità, 1972
––., S. D'Amato, Epigrafia della regione dei M., 1975.
Src: de Vido, Stefania (Venice) and Wiegels, Rainer (Osnabrück), “Marsi”, in: Brill’s New Pauly, Antiquity volumes edited by: Hubert Cancik and , Helmuth Schneider, English Edition by: Christine F. Salazar, Classical Tradition volumes edited by: Manfred Landfester, English Edition by: Francis G. Gentry. Consulted online on 30 May 2019 http://dx.doi.org/10.1163/1574-9347_bnp_e725090
First published online: 2006
First print edition: 9789004122598, 20110510
Now after discarding a late mythological inventive addition, the actual etymology is quite complicated to ascertain in al the nuances.
After comparing Georges Dumézil's "Archaic Roman Religion" and "Camillus" with
To demonstrate this requires a slight excursus on the fate of -rs- clusters in Sabellic. As so often in Sabellic, we must recognize two different treatments depending on whether the cluster is original or arises by medial syllable syncope. Original *-rs- clusters are retained in Umbrian: tursitu < *torsētōd ‘frighten,’ çersiaru (name of a month) < *k̑ ersii̯o- ‘pertaining to cutting,’ i.e., the month of harvest (cf. Hom. Gk. ἀ-κερσε-κόμης ‘with unshorn hair’).9 But in Oscan the *s is lost with compensatory lengthening of the preceding vowel: teer[úm] ‘earth’ (CA A 12) < *tersom. N.b. the spelling of the long vowel with geminatio in an inscription in the reformed Oscan alphabet, which shows that the long vowel is secondary and not to be directly compared with OIr. tír ‘earth.’ Secondary *-rs- clusters (rectius *-rz- clusters) arising by medial syllable syncope become -rf- in Umbrian but assimilate to -rr- in Oscan: e.g., Umb. parfa (type of bird) < *parasa- (cf. Lat. parra,10 but Osc. kerrí ‘Cereri’ (TA A 3) < *keresēi̯).
Generally Paelignian is a kind of north Oscan (with a South Picene admixture) and thus we should expect it to behave like Oscan in regard to its treatment of -rs- clusters. And in fact the outcome of secondary *-rs- is well attested in Paelignian in the adjective Cerria ‘of Ceres’ (Pg 18) < *keresii̯o-.11 The unavoidable conclusion is that cerfum cannot be related to Ceres since neither an original nor a secondary *-rs- cluster will yield Paelignian -rf-.12 The form cerfum, on the other hand, cannot be disassociated from the Umbrian god Çerf(o)- and this means that this name too cannot be related to Ceres. The only possible sources that will give a Paelignian and Umbrian -rf- are sequences of r plus a labial or dental voiced aspirate.
Michael Weiss: "An Italo-Celtic Divinity and a Common Sabellic Sound Change", Classical Antiquity, Vol. 36, Issue 2, pp. 370–389, 2017. (DOI)
It becomes clear that all the variants of the god that map to modern single word Mars are quite numerous, and indeed used frequently to denote Italian people, albeit not necessarily to be interpreted as the received war-god, more akin to Apollo as just an organiser and patron.
According to Festus (116,2; 150,34), the Oscan form of Mars. The appearance of M. in Oscan dedicatory inscriptions (Vetter 196; [1. no. 177, 179]: 3rd/2nd cents. BC) and the Oscan roots of the Mamertini, important since the 4th cent. BC, seemed to support Festus [2. 155, 167, 172]; this led to the marginalization of Varro's postulate of the Sabine origin of M. (Varro, Ling. 5,73). The so-called Lapis Satricanus (AE 1979, 136), found in Satricum 50 km south-east of Rome, a dedicatory inscription Mamartei (‘for Mamars), is proof of the existence of a Latinized form at c. 500 BC. It remains doubtful, however, whether this means that Oscan and Sabine M. are derived from a Latin-Roman Mars [3. 293-295] and that the confusing variety of other dialect forms originated here. It may be, instead, that Mamars in Satricum is the result of linguistic influence from the Sabine [4. 85-87].
Phillips, C. Robert III. (Bethlehem, Pennsylvania), “Mamers”, in: Brill’s New Pauly, Antiquity volumes edited by: Hubert Cancik and , Helmuth Schneider, English Edition by: Christine F. Salazar, Classical Tradition volumes edited by: Manfred Landfester, English Edition by: Francis G. Gentry. Consulted online on 30 May 2019 http://dx.doi.org/10.1163/1574-9347_bnp_e719940 First published online: 2006 First print edition: 9789004122598, 20110510