Speak / Fala – Freeing Man-Machine Conversation


What sound do we hear when the machines “talk” to each other? We are partly excluded from this conversation because the human body is not designed to “hear” machine language. Through the work “Speak / Fala” Rejane Cantoni and Leonardo Crescenti attempted to make machines talk our language in order for them to communicate among themselves and with us. The installation “Speak / Fala” consists of a small crowd of mobile phones placed on microphones stands. Triggered by the sound of a human voice that pronounces something, the phones start to engage in a game of Chinese Whispers, taking it in turns to pass on the message. But in what sense do they “talk”? They emit words in twenty different languages ​​(the number of languages ​​understood by the devices chosen). When someone speaks a word into the microphone, the first device repeats using the language on which it is set, choosing the most similar word to the one heard from the internal vocabulary of about 3000 entries. This similarity can be at the semantic level, if the device is able to interpret a connection. The connection can also be phonetic, according to the sound similarities between different languages. Of course, the last word in the chain will not necessarily be the one started with. In the words of the artists, most of these conversations turn out as “waterfalls of words” joined together by a kind of poetic link switching between sound and meaning. Through a harmonic sound stream that could be the signifier or the meaning, the work seems to remember the unique matrix of our languages. This harmony is actually created by the algorithmic and voiceless machine language, which silently remains the only universal language currently possible.
Chiara Ciociola

Speak / Fala

SPEAK from cantoni-crescenti on Vimeo.