Microcontact. Language Variation and Change from the Italian heritage perspective is an ERC-funded research project that aims to understand language change from a syntactic microvariational perspective, by observing core syntactic aspects of Romance languages in contact.

Read more


Microcontact. Language variation and change from the Italian heritage perspective

1. Introduction

At the turn of the 20th Century, many Italians emigrated to the Americas. In doing so, they created one of the largest real-life laboratories for linguistic research. The ERC-funded Microcontact project, which studies variation and change from the Italian heritage perspective, will use this hitherto unexplored laboratory to understand how languages change.

Italian emigration was a massive phenomenon. In 1916, the year of the largest migration, 872.598 Italians left Italy (source: ISTAT, Italian National Institute for Statistics). After World War II, a new wave of migration took place: according to the ISTAT, around 400.000 people left Italy between 1950-1960. Interestingly, most of these Italians did not speak Italian as their native language, but some “dialect”. With this term we traditionally refer to those Romance languages spoken in Italy, having evolved from Latin, which are sister languages to standard Italian. Many of them, like Neapolitan, Florentine or Venetian, were official languages of some important trade, cultural or political center for centuries, and as such have a venerable written tradition. Although they are referred to as dialects, they are fully fledged (Romance) languages, with their own grammar and their own lexicon.

When these Italian emigrants moved across the Atlantic, their languages entered in contact with other Romance varieties, like Argentinian Spanish, Brazilian Portuguese, or Québécois French, as well as with English. Meanwhile, Italian in Italy started expanding and being learnt by all social classes, and hence after World War II these languages came to a more extensive contact with Italian as well: an important part of the laboratory can be found in Italy.

The languages spoken by all of these first-generation emigrants potentially provide a unique window into the mechanisms of language change in general, and of syntactic change in particular. The project refers to these languages as “heritage” in a broad sense, and considers in particular first-generation speakers.

A theory of language change

There are very few other cases of speakers getting into intensive contact with such an assortment of closely related languages. We know that languages tend to change when in contact with other languages. Unfortunately, aside for some sporadic collection of data, mainly performed for cultural studies, we do not have any systematic documentation of what happened to Italian heritage languages in contact. According to some studies, first generation emigrants (i.e. those emigrants who moved to the Americas after WWII and are still alive) present significant attrition (i.e. decline of L1 competence and proficiency, Lambert & Freed 1982) and shift (i.e. decline in usage of L1 and increased use of L2) (Di Pietro 1960, 1976, Saltarelli 1986, Simone 1988, Haller 1993 a.o. for Italian languages).

These studies mainly focus on the sociolinguistic situation of contact, and on lexical borrowing, but not on changes in the grammar.

Grammatical change is one of the most complex linguistic phenomena: there is no apparent reason why languages should change. Yet, they do, sometimes even despite systematic socio-political efforts to prevent this change.

It is usually assumed that language change can happen in two ways: it can be spontaneous, that is ‘endogenous change'(EC), when something in a grammar changes without any external cause. Alternatively, change can be contact-induced (CIC), when it is caused by contact between two or more languages. One of the main problems when dealing with language change is that it is almost impossible to ascertain what has caused it at a given stage (Weinreich, Labov & Herzog’s 1968 actuation problem). We have some reasonable hints in some cases, but in most cases identifying its cause is extremely difficult, if not impossible. Also, even if a language A is in contact with a language B, it is not obvious that any change occurring in A is due to the influence of B.

In this project, we aim to develop a theory of syntactic change in contact, by observing how specific syntactic structures react to both EC and CIC. Italian heritage languages offer a unique combination of wide diachronic written documentation and multiple contact with other, minimally different, languages, which will allow us to investigate EC and CIC and integrate syntactic theory with the tools that are necessary for their analysis.

The study of EC and CIC usually examines two stages of a language: Stage1- before the change -, and Stage 2- after the change-. This project introduces a third observation point: the “in between” stage, which will be provided by microcontact. With microcontact here we mean contact between two minimally different syntactic systems (grammars). While microvariation will make parts of grammars comparable and will make it possible to have “snapshots” of languages while change is happening, genetic and typological similarity between the languages in contact will allow us to control for one element at a time: the languages in contact do not differ very much since they are genetically related.

This project aims to give an answer to the following major research questions:

1. How does contact-induced syntactic change (CIC) happen?

2. What are the main differences between EC and CIC?

3. Are there elements in the grammar that are more prone to contact-induced change, and are there elements that are more prone to spontaneous change?

1. Documenting heritage Italian languages

The Italo-Romance heritage varieties constitute an exceptional lab for this study. We will examine first-generation speakers, i.e. those that moved to the Americas between the 1940s and the 1960 (as well as their peers who stayed behind at the time and came into extensive contact with Italian).

The documentation of these heritage languages is scarce, and quite fragmented. We know very little of the language of the first-generation emigrants. We therefore need to conduct a large-scale documentation effort before we can start our analysis. For this task, we will make use of a crowdsourcing software, addressed to speakers of heritage languages, who will be involved for the first time in active scientific research about their own language.

We will contact young generation speakers through embassies, consulates and associations, requesting that they record their grandparents and upload the recordings on an interactive atlas. This will help us identify those communities and speakers that best suit the research profile that we need, whilst simultaneously involving speakers in the documentation and research on their language. The data documentation will continue with extensive fieldwork in these communities.

Specifically, we will look at 4 Italian dialect groups (2 languages per group) in diachrony and in contact, as illustrated in Figure 1, illustrating the varieties that will be considered, at different stages of their diachronic development and in contact with the languages in America.

Figure 1.

2. Documenting Heritage Italian Languages

Our main hypothesis is that EC and CIC are different in at least two ways. Roughly, speaking, when languages are just left on their own (in situations of EC), they follow very specific paths leading from one type of grammatical construction to another. This path is however disrupted in CIC, where ‘jumps’ can occur in the change. These jumps may also lead to instabilities elsewhere in the grammar.

In order to identify the mechanisms of change in progress, we will focus on optional structures. With optional we mean here two perfectly interchangeable syntactic structures within one language. The term optionality is quite problematic in a deterministic theory like generative grammar, because in this view a grammar will give you precisely one way in which to express a given meaning, and not two or more.

3. Theoretical Issues

According to Kroch (1994), syntactic doublets should be categorically excluded from stable grammars, although they can be attested in systems that can be viewed as unstable for one reason or another, e.g. developing and interlanguage systems, dying varieties and those undergoing change more generally. These doublets are exactly those that we will be searching for, in order to observe change in progress. Optionality will hence refer to syntactic doublets in systems undergoing change, be it EC or CIC.

In addition to optionality, also the notion of markedness plays a crucial role in our theoretical investigations. marked structures, we assume, are structures that specify the same information more than once. For example, sentence (2), where the subject is expressed twice (with Maria and with la), is more marked than sentence (3), where it is expressed only once.

(2) La Maria          la                  magna  [Venetan]

     the Mary-subj she-subj          eats

     ‘Mary eats’

(3) Maria         mangia            [Italian]

   Mary-subj     eats

   ‘ Mary eats’

We can establish hierarchies of markedness, because different forms can express more and more information. (For example, in the examples above, the subject is still partly expressed on the verbal ending –a; it would be even less marked to leave this out.)

This hierarchy can be expressed in different ways, and on different elements (syntactic structures, syntactic features, etc).

Our hypothesis is that EC will follow a path from more marked to less marked, or from less marked to more marked, one step at a time. CIC will be able to tackle any point of the hierarchy instead: it just depends on how the ‘other’ language is. For heavily marked constructions, this means it will be possible for them to become simpler, or disappear, or skip one or more steps in the hierarchy.

4. Empirical Issues

As an example (out of the several we study in this project), consider ‘auxiliary selection’. In English, the perfect tense of verbs is always made with the same auxiliary HAVE: I have come, I have eaten. However, in many languages in the world, different verbs use different auxiliaries, often either be or have. For instance, in Italian, one says Sono venuto ‘I have come’ (lit. ‘I am come’) and Ho mangiato ‘I have eaten’. Many upper southern Italian dialects present a very interesting system of auxiliary verbs, whereby the choice of subject of the sentence determines the choice of the auxiliary. We exemplify the problem through Abruzzese, but a similar phenomenon is found in several varieties of Neapolitan:

(4)  1st and 2nd person singular and plural: BE

      3rd person singular and plural: HAVE

(5) a. So, si, seme, sete                             liggiute nu libbre/       durmite/           partite  [Abruzzese]

          be-1st.sg, 2nd.sg, 1st.pl, 2nd.pl    read     a book             slept                left

          ‘I, you, we, you have read a book/slept/left’

b.  A                               liggiute nu libbre/       durmite/           partite

have-3rd.sg=3rd.pl       read    a book             slept                left

‘S/he, they have read a book, slept, left’

(4) and (5) show that in these languages not only is the ending of the auxiliary different depending on the subject, but also the root is. We can define this as a sort of subject doubling phenomenon: the information that we are dealing with a second person is expressed both on the s (=be) in the auxiliary part and in the second person singular ending –i. According to our definition, this is therefore a marked structure. The auxiliary has agreed with the subject twice, as it were.

The same varieties (with some exceptions) put  the preposition a before the direct object of a sentence, but this happens only if the object is (animate and) 1st or 2nd person. This process is called Differential Object Marking (DOM). Note that in Romance this preposition a usually marks animate objects only, thus person-driven marking of the object is also some sort of doubling, marking both animacy (because the speaker and addressee are of course animate) and first or second person. An example of DOM is in (6):

(6)  So viste a mme/ a tte/ a nu/ a vu / *a jisse/*a Marije/*a esse                                         [Abruzzese]

      am seen to me / to you/to us/ to you/ to them/ to Mary/ to her

      ‘I saw myself, you, us, you, them, Maru, her’

We will investigate what happens when each of the two phenomena get in contact with:

  • Argentinian Spanish – no auxiliary selection (no use of the present perfect, thus no auxiliary for the past tense) / animacy-driven DOM

  • French and Italian– argument structure driven auxiliary selection  / no DOM

  • Portuguese – argument structure-driven auxiliary selection, but different auxiliary (ter instead of have/haber) – partial DOM

  • English – no auxiliary selection (only HAVEhave used for the present perfect) / no DOM.

Furthermore, we know from previous research (D’Alessandro 2014) that in Italy the Abruzzese auxiliary selection system is changing, in the direction of an expansion of have. Without more data, we cannot ascertain whether it is a spontaneous change or it is induced by the contact with Italian. Furthermore, we need to investigate whether the person-driven system of auxiliation and of DOM is an innovation, as we think. This would mean that some new information was introduced at some point in the history of these languages: the person-driven selection, and hence that the construction became more marked. By observing Abruzzese in contact with French, Spanish, English and Portuguese, we expect the following:

  • if change is EC, heritage Abruzzese in contact with French, Portuguese, Spanish and English will most likely go in the direction of extending have, as has happened in other Italian languages, like Sicilian, i.e. towards markedness reduction (no reference to the subject of the sentence in the root, but only in the ending). The structure of the contact languages will have no impact. As for DOM, it will remain as is.

  • if change is CIC: there will be a different behavior depending on the contact languages; there could be a total drop of auxiliary selection in heritage Abruzzese varieties in contact with English and Spanish, for instance, while the extension of have could only be found in contact with French. The contact with Portuguese, which has auxiliary selection but with different auxiliaries, will be the most telling. One possibility is that aux selection will be reinforced, with the adaption of new auxiliaries; another that it will be dropped. The same applies to DOM: different contact languages should have a different impact on person-sensitive DOM. The contact with Spanish might induce the loss of the person-driven specification, to leave only the animacy DOM.

Person-driven auxiliary selection and DOM are of the phenomena that can tell us about change, and that we wish to investigate within this research. Others are deictic adverbs, adjectives and pronouns in southern Italian languages and subject clitics (and possibly negation) in northern Italian ones. All these structures are heavily marked, and will allow us to obtain a detailed picture of the mechanisms of change by looking at their evolution as EC (in diachrony) and CIC (in microcontact). Again, each pair in contact alone won’t tell us much: it is all pairs together that compose the picture of change, like tiles of a gigantic puzzle.

5. Organizing the work

This project involves a part of diachronic microvariation, which serves to assess whether the phenomenon has been relatively stable during the last 5-600 years, and a part of synchronic microvariation in contact. The languages investigated are illustrated in Figure 2.

Figure 2.

The project includes 3 Phd subprojects (one for person-driven auxiliary selection and differential object marking in contact; one for deixis in contact; one for subject clitics in contact), and one postdoc working on the diachrony of these constructions. The principal investigator is responsible for the synthesis of the theoretical and empirical findings, as well as for developing a theory of contact-induced (micro-)change. The project will target four main Italian linguistic macroareas: North, Tuscan, Upper South, Extreme South.

Two languages per group will be addressed, both microdiachronically and in a micro-contact situation:Venetan and Piedmontese for the northern group (which, as you can see from the map, is rather varied; the phenomenon we will address, subject clitics, is common to both); Florentine and Sienese for the Tuscan group; Neapolitan and Abruzzese/Teatino for the Upper southern group; Palermitan andSalentino for the extreme southern group. The choice of Florentine is obviously linked to the fact that this vernacular was the basis of modern standard Italian, and as such it was used in most written documents of the Italian peninsula. While all other varieties have had some formal status and are largely documented historically, coastal Abruzzese does not have a long literary tradition. However, a large number of studies has been dedicated to this language by the PI.

Each of these 4 groups mentioned above will be checked in contact with the following varieties: Argentinian Spanish; Brazilian Portuguese; Québécois French, and US English (as a control variety). Furthermore, contact with Italian in Italy in the last 60 years will also be addressed.  Ideally, it should be possible to check three or four contact points for each variety. These contact points will be selected based on the availability of speakers who will agree to participate in the project, as well as depending on the phenomena at issue. Giving the vastness of the research areas, each team member will be responsible for one area. The data collected via crowdsourcing as well as via fieldwork will be available to all team members.

The full project can be found here: ERC_CoG_2015_B1+B2_Final