earliest Luwian seal
The Dutch Anatolianist Fred C. Woudhuizen has published a discussion of a dated seal with hieroglyphic signs on it which belongs to c. 2000 BC and so becomes the oldest record of any Indo-European language. The previous claimant was fragments within Akkadian texts at Kanesh, so loanwords and personal names which later information allows us to fit into a context and to identify as Hittite. The seal was found at Beycesultan, in south-west Anatolia, inland but probably developed (and literate) as the hinterland of an Aegean coast which was the site of very vigorous trading activity and cultural stimuli. The site has been identified with ancient Mira. It is on the upper reaches of the river Meander.
The seal was originally recovered in 1958.
The paper is “Stamp Seal from Beycesultan”, 2012, and I downloaded it from the Internet.
This changes the history of the Indo-European languages. The object belongs to an era which is overwhelmingly silent, and must be atypical; but it is not itself silent. It is the first sound of a whole language group.
The oldest IE now clusters geographically: Luwian, Hittite, Cretan writing in Mycenaean Greek.
Incidental points are as follows.
The previous “oldest IE” known was not before 1950 BC.
Early written material from Western Anatolia is rare and we would expect the very earliest inscription to be in the south-east, close to the origins of writing in Mesopotamia.
The hieroglyphic writing system is original, and we would have expected something in cuneiform, imitating the thriving scribal industry of Mesopotamia and Syria, to get there earlier.
This very brief inscription has some signs from a syllabary. However, the phonetic values of the signs may be different from (more archaic than) the Luwian language as we know it from inscriptions mainly a thousand years younger. The seal says "(commander of) a thousand", giving the title of the owner. This corresponds with numerous other Luwian seals. The meaning is "the chiliarch of Mira (over) the river and this town", roughly.
The seal is precisely dated by stratigraphy. It was found by famous archaeologist James Mellaart. There is very convincing material, in print and on the Internet, about Mellaart’s fantasy material. It seems that he drew fabulous objects (wall paintings etc.) which did not exist, but there is no trace of him faking stratigraphy or forging objects. In fact, he did not recognise the scratches on the seal as forming a written message. Mellaart complained (in 1995) that in several years of digging at Beycesultan in the 1950s they had found no texts at all. He found the seal in 1958 and it was not seen as writing until the 1990s. So it is not very plausible that Mellaart faked it, and we can rely on the genuineness of the seal.
The Luwian seal-script bears no resemblance to the Cretan hieroglyphic A seal script. The two scripts may have originated in neighbouring regions, and be linked to an Aegean trade economy which involved both regions.
The Luwian hieroglyphs in their earliest form look very stylised and mature, that is, old. They are already far removed from pictures.
I do not know of any reason why the hieroglyphic script (which is a syllabary, not solely ideograms) thrived in parallel to cuneiform, or why the Luwians abandoned cuneiform in the written records of their late kingdom around Carchemish. There are ample Luwian records in cuneiform in the Hattusa archive (and this is a great help in reading the hieroglyphic version of their language).
The seal script is thinly recorded before about 1500 BC. It is known from sealings as well as from the seals.