24. September 2020
The ERC project Microcontact investigates language change in contact from a heritage Italo-Romance perspective. We performed extensive fieldwork in Argentina, Brazil, US, and Canada and looked at the way the language of the speakers of Italo-Romance varieties (i.e. all those Romance languages spoken in Italy that are not Italian, e.g. Venetan, Sicilian, Neapolitan, etc) have changed in contact with Argentinean Spanish, Brazilian Portuguese, Quebecois French, and English (as control). We call contact between languages that belong to the same “macrogrammar” microcontact.
We collected around 500 spontaneous conversations and about 300 tests (forced choice or completion tasks). Our speakers are 1st generation emigrants (rather elderly, between 75 and 90 years old) their children (heritage language speakers) and grandchildren.
These are the main findings so far:
- it is usually believed that the omission of the subject by heritage language speakers declines because that requires too much competence in syntax but is also too costly in terms of discourse resolution and memory load. Most heritage language studies report the insertion of overt subjects where baseline grammar speakers would drop them.
Our study found that subject drop is not affected at all by contact between languages that are typologically very similar or genetically related, like in our case.
- it is usually believed that the special marking of the animate object in a clause (called Differential Object Marking) weakens in heritage speakers. We found the opposite: in fact, an increase of DOM.
- we also found that indexicals (like pronouns, or demonstratives) do not seem to be affected at all in microcontact.
We cross-checked these results with creole languages (a different form of contact) and we found that those creoles that have these phenomena behave like microcontact languages, not like “normal” contact.
In order to identify the key factor involved in language “stability” we tested objects in their “natural” position and in a prominent, topic position, and we found that indeed topicality (i.e. expression of prominence in discourse) and indexicality (i.e. expression of reference to an “object” in the world) are the relevant stability factors. Topicality and indexicality both have underlying the function of “linking” what is being said either to what was said before or to the world.
Our conclusions are that when speakers have no perception of the difference between the parts of grammar involved in contact, either because they are dealing with two “dialects” of the same macrogrammar, or because they are dealing with too many, too different, grammars (like in the case of creoles), they resort to cognitive strategies to determine their grammar. In our case, linking is the cognitive key. These strategies are fundamentally different from grammar-internal mechanisms, which are more readily resorted to when the speakers perceive the difference between two clearly distinct grammars in contact.
19. November 2020
Phonetic convergence, defined as an increase in similarity of the
speech patterns of two interlocutors in communicative interaction, has
been studied thoroughly in human-human interaction. There is now an
increased interest to explore phonetic convergence in human-computer
interaction as well, as spoken dialogue system are becoming integrated
in our everyday life.
In our recent work we have demonstrated that human experimental
subjects show patterns of phonetic convergence when they are exposed
to synthetic voices in a shadowing experiment, similar to the
convergence patterns observed in human-human interaction. In ongoing
work, we are extending the experimental approach to a Wizard-of-Oz
scenario, to examine intonation of wh-questions and pronunciation of
allophonic contrasts in German. We find that almost every experimental
subject converges to the system to a substantial degree for a subset
of pertinent features, but we also find maintenance of preferred
variants and even occasional divergence. This individual variation is
in line with previous findings in accommodation research. In another
experimental setting we are examining how the presence of other
speakers affects the interaction with a spoken dialogue system, in two
conditions: with and without additional speech input from a human
confederate as a third interlocutor. The comparison was made via tasks
performed by participants using a commercial voice assistant under
Ultimately, we aim to develop a quantitative model of phonetic
convergence in spoken human-computer interaction, build synthetic
voices that are capable of adapting their speech output to the user's
speech patterns, and derive implications for the design of
conversational interfaces in speech technology.