Our orthography proposal: Simple Kostakis

The Tsakonian language, the last remaining descendant of the Doric Greek spoken by Spartans in Ancient Greece, is at a crossroads: after more than 2 500 years of history, it is in critical danger of extinction with less than a few thousand speakers left and limited transmission to younger generations. In 2023, the Tsakonian Digital project was launched to help preserve this heritage and foster its usage by the development of digital tools and AI models. Despite the new possibilities opened by modern technologies, two paramount challenges hinder its development in a digital era: both scarcity and lack of homogeneity of Tsakonian written data due to the lack of consensus in the orthographic rules used. But that is not all the story.

It turns out that Tsakonian does have a well-established and suitable orthographic norm, developed by the Tsakonian linguist Thanasis Kostakis. Taking the Greek alphabet as a basis, it adds several marks on various letters to represent sounds absent in Standard Modern Greek, such as τ̔ (vs. Modern Greek τ) to represent an aspirated t sound, among others. While accurate and tailored to the language, major computer and mobile systems do not incorporate these symbols in the Greek keyboard. This prevents regular users from employing it, reducing the already limited pool of potential Tsakonian language users in a digital environment. Consequently, some speakers have adopted their own personal solutions by using combinations of letters to represent these sounds alongside other orthographic rules.

These circumstances directly impact both data quantity and quality, since converting different orthographies to a common standard requires extensive effort, increasing the amount of time and effort needed to develop digital tools. It is precisely these tools what would help the language survive in a digital setting and prevent local speakers from switching to other languages out of necessity. This renders the orthographic consensus one of the major open topics in Modern Tsakonian yet to be resolved.

The simplest solution would be to add the Tsakonian keyboard to mainstream devices with no or little configuration to lower the entry barrier to write using the Kostakis orthography. However, given the limited number of speakers and commercial interest in the language, it is not to be expected that this will be implemented soon.

In the meantime, the Simple Kostakis orthography provides a straightforward, ready-to-use and digital-friendly solution. By preserving Kostakis' precise and consistent rules, but removing special orthographic marks, this system allows typing Tsakonian using only the characters contained in the Greek keyboard without extensions. To do so, it retains some commonly used digraphs and introduces some new ones:

• τθ for [tʰ], an aspirated t
• κχ for [kʰ], an aspirated k
• πφ for [pʰ], an aspirated p
• ζζ for [ʒ], as in French j'aime
• ξξ for [ʃ], as in English shop
• τσσ for [/t̻͡s̪/], as tz in Basque

Following Kostakis' orthography, it also distinguishes the following palatalized pairs:

• νι [ɲi] vs. ν'ι [ni], such as Spanish ñ [ɲi] vs. n [ni]: moño vs. mono
• λι [ʎi] vs. λ'ι [li], such as in Peninsular Spanish without yeism: arollo vs. arroyo

After some initial training, this set of rules allows speakers to write the language without any external tools, fostering the creation of much needed Tsakonian written data. Furthermore, if in the future consumer mobiles and computers add support for a Tsakonian keyboard with special marks, users can seamlessly transition to the regular Kostakis orthography with no effort. This proposal aims to help closing this longstanding debate considering the language's current circumstances and technological context. Further ideas on how to simplify and implement this proposal are more than welcomed.

And all this means that, while AI can perform much of the heavy lifting for us, it is still up to us to create the right conditions to do so.