Inscriptions as data: digital epigraphy in macro-historical perspective

Abstract

Even the exact number of ancient inscriptions, dating between 800 BC and 800 AD, currently estimated at 600,000 surviving Latin and Greek ones, remains elusive due to its fragmentation to multiple sources (Bodel 2001, 4; Beltrán Lloris 2015, 135).Although numerous analogue and digital corpora exist, their contents and structure overlap to a variable degree and are not straightforward to combine, providing a major obstacle to comparative studies.While the challenge of technical data wrangling and distant-reading the inscriptions may be beyond a single researcher, a collaborative team equipped with both the necessary computational tools and domain-speci c skill set may get further.Working from established digital repositories and building our digital toolkit on the go, we explore the spatiotemporal variation in the distribution of surviving epigraphic monuments and assess the dependency of comparative research on the state of modern-day research infrastructure. Epigraphic habit(s) 40 years later: can we do any better?Inscriptions vary signi cantly in form and contents, touching almost all aspects of life in an ancient society.Commissioning an inscription was part of a long-term cultural practice that served as a physical response to a change of circumstances, i.e. death, birth, change of legal status, social rank, physical appearance or they can serve as a reminder or re ection of historical events.The inscribed texts range from short private messages on objects of daily use to extensive regulations and codes issued by a political authority, to monumental inscriptions helping to create a public image and build social prestige.In contrast with literary texts, which survived via multiple copying and editing, the text of inscription has the original unaltered form with original typographical and grammatical errors, being an ideal study subject for historical linguists and historians.The information preserved via inscriptions has potentially broader demographic coverage than literary texts, as social groups producing inscriptions were of more diverse socio-economic and cultural background than those of literary texts (Bodard 2008).Similarly to archaeologists who are never sure they have all the pots, epigraphers will never have access to all the inscriptions ever created.Duncan-Jones (Duncan-Jones 1982, 360-362) has estimated a rate of preservation for one type of honori c inscription in the cities of Northern Africa at 5 %, but some consider his gure too high and only relevant to the local area (Noreña 2011, 181).Should the 5 % survival rate apply and be uniform throughout the Mediterranean, then the 500,000 surviving Latin inscriptions would be a sample of 10 million original inscriptions.Should the rate be 1 %, then we are missing some 50 million inscriptions.These remain guesstimates as the factors in uencing the survival of inscriptions, such the availability of durable material, the completeness of archaeological monument surveys, or even the level of plundering or secondary use of inscriptions have not been uniform throughout the Mediterranean (Mann 1985, 204-206; Keppie 1991, 34).The corpus of surviving inscriptions has its biases and is far from a carefully strati ed sample.The haphazard process of survival resembles simple random sampling, a method used in quantitative research to lower the margin of error.While a random sample may not be useful 10 11 12 13

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 42 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.