Readability assessment has been a key research area for the past 80 years, and still attracts researchers today. The most common measures currently (2011) in use are Flesch-Kincaid and Dale-Chall. Traditional models were parsimonious, incorporating as few linguistic features as possible, and used linear regression to combine two or three surface features. Later models used psychological theory, measuring such things as coherence, density, and inference load. A variety of machine learning models were used and one neural network. Key surface linguistic features were average syllables per word and sentence length. The Machine Learning methods performed well. Machine Learning methods can improve readability estimation. The process is data-driven, requiring less manual labour, and avoiding human bias. Current research seems to focus on deep learning methods, which show great promise.

Many have expressed concerns regarding the replicability of scientific research. However, little of this ongoing discussion has focused on research examining the production of vowels and consonants or the many important choices that researchers must make in pre-analysis phases of speech production research. The literature reviewed here indicates that not all speech production studies have been replicated, and that how speech is elicited may affect the results that are obtained. Many different elicitation techniques are in current use, but none represents a gold standard. The new Characteristic Speech Production (CSP) technique presented here aims to augment replicability by obviating the need for participants to accommodate their speech to that of others or adopt a particular speaking style as they give meaningful answers to meaningful questions. Given the novelty of the CSP technique, the chapter provides a protocol that is designed to test its efficacy. If the CSP technique can be shown to yield speech samples that are more representative of individuals’ speech than a standard list-reading technique, a change in how speech is elicited for production research will be warranted.

It has been widely recognized that how languages behave, particularly under conditions of contact with other languages, depends on their context. Using the Ethnolinguistic Vitality framework, this chapter describes the demographics, linguistic attitudes and institutional supports for heritage languages, defining the concepts and illustrating them with examples from Toronto, the context in which the HLVC project is conducted. Demographic information includes population sizes, language shift rates, and history of settlement in Toronto. Status information includes both reflections on the status of heritage languages, as a whole, in Canada and labels attributed to the specific varieties. The institutional support section reports on the number of language classes available for each language. The chapter also includes discussion of language policy, particularly for education, and the demographics of the university where the research is centered, enabling other researchers to best consider what aspects of the project might need adjusting for adaptation in other contexts.

This chapter reports on trends of continuity and divergence within the heritage generations examined and between heritage and homeland varieties. It discusses the degrees of similarities between the varieties in terms of (a) rates of use of innovative forms and (b) conditioning factors in the constraint hierarchy. The three variables examined are voice onset time (VOT, n=8,909), case-marking on nouns and pronouns (CASE, n=9,661), and variable presence of subject pronouns (PRODROP, n=9,190), each in three or more languages. The similarity in rates and conditioning effects across generations for (PRODROP), examined in seven languages, particularly contrasts with findings for this variable in experimental paradigms. Similarly, findings of little simplification or overgeneralization of the case system in three languages stands in contrast to the outcomes of several previous studies. (VOT) shows a drift toward (but not arriving at) English-like values for only some of the languages examined. For each variable, models are presented and interpreted; a table then details which aspects of the analysis contribute to the interpretation of stability and of each type of variation.

The variables examined in Chapters 5 and 6 show little evidence of being used for identity work. That is, they do not show (consistent) effects of ethnic orientation measures or speaker sex. This chapter explicitly contrasts variables that reflect indexicality (correlation to social factors) in homeland varieties to non-indexical variables. We begin by considering three indexical variables in Italian: (VOT) in unstressed-syllable contexts, (APOCOPE), and (R), illustrating the extent to which indexicality is maintained in the heritage variety. We find increasing use of the more standard variant only in (VOT). Furthermore, we find that younger speakers (both in homeland and heritage) favour the non-standard variant. We then compare the variable (R), the contrast between trill (or tap) and approximant variants, in Italian and Tagalog, where it has indexical value in the homeland varieties, to Russian and Ukrainian, where it does not. Finally, we consider two additional indexical variables: Cantonese denasalization and Korean VOT. We conclude by contrasting the behavior of homeland-indexicals in heritage varieties. The presence of indexical value in homeland varieties does not consistently influence outcomes in the heritage varieties.

This chapter draws cross-linguistic comparisons among the patterns reported in Chapter 5 for three linguistic variables that occur in at least three languages in the project: (VOT), (CASE), and (PRODROP). Conditioning factors, both linguistic and social, are discussed. Collapsing across rate and constraint hierarchy for each variable, we note any indication of change in either. Half the context we examine exhibit stability. Of the eight that indicate difference, half of these can be attributed to English (including both convergence and divergence). With few differences between homeland and heritage speakers to work with, we find few generalizations about what parts of the language, or which languages, change. We do see more change in one morphosyntactic variable, (CASE), than in the phonetic variable (VOT), but less in the other morphosyntactic variable (PRODROP).

The chapter surveys repetitions and reduplications in Italian, from the segmental to the discourse level. Italian has reduplicative structures in ideophones, onomatopoeic formations, child language, and baby talk; segment repetition is used as an expressive device in commercials and product names; reduplication is used as a lexeme formation device in Verb-Verb compounds such as fuggifuggi ‘stampede, lit. run away run away’, and as a means of intensification of adjectives and adverbs; some sequences of two nouns have lexicalized with adjectival or adverbial meaning; contrastive focus reduplication is also attested in Italian. Discourse markers are often reduplicated; several cases of repetition of imperatives in discourse have constructionalized, giving rise to converbs with concessive or hypothetical meanings or used as antecedents of consecutive clauses; noun reiteration in discourse can be used to indicate frequency of occurrence of entities and events. It is argued that no clear dividing line can be drawn between pragmatic or syntactic repetition and grammatical or morphological reduplication, since grammaticalization of discourse repetition in diachrony often occurs.

Affix counterposition has two subcases: in the first (exemplified by the inflection of reflexive verbs in Lithuanian), an affix that is suffixed to the stem in some words is suffixed to a prefix in others; in the second, mirror-image case (exemplified by Noon adjective concord), an affix that is prefixed to the stem in some words is prefixed to a suffix in others. Rule aggregation models the phenomenon of affix counterposition as a deviation from the stem operand criterion involving a rule of affixation R whose operand is ordinarily a stem (when R isn’t aggregated) but is instead an affix (when R is aggregated to the rule introducing that affix). Rule aggregation brings affix counterposition into conformity with the affix directionality criterion. Nevertheless, there are real deviations from the latter criterion: some languages have true ambifixes that actually function as prefixes in some word forms but as suffixes in others. Gurma noun-class inflection exemplifies this possibility. Moreover, the morphotactics of Italian pronominal affixes involves a significant interaction between true ambifixation and rule aggregation.

The quotative system is routinely adduced as the locus of rapid cross–linguistic change. Aside from the prodigious number of empirical studies investigating English quotatives, quantitatively driven demonstrations of change in the quotative system of other languages remain the exception to the rule. Observing that change in languages other than English has often been intuited from isolated or anecdotal examples, we inaugurated a large–scale study of quotative variation in European and Canadian varieties of French, supplemented by data from Brazilian Portuguese and Italian. Drawing on more than 5,500 tokens representing the targeted varieties, detailed quantitative investigation revealed that only in urban varieties of Quebec and Acadian French does the innovative être comme variant (cf. English be like) qualify as a mid–range – and locally conditioned – change in progress. In other varieties that we examined, including the French and Portuguese spoken in the global cities of Paris and São Paulo respectively, we find little compelling evidence of anything other than relatively incipient change in the quotative system. Taken together, our quantitative results are damaging to ubiquitous claims that simultaneous parallel developments are purportedly affecting the quotative system of numerous languages and point to the primordial importance of community–based speech data in ratifying linguistic change.

The evolution of the Romance languages from Latin was significantly shaped by the numerous language contact environments, which resulted from conquest, colonization, and trade. This chapter traces the development of the largest Romance languages throughout Europe, with emphasis on the known or postulated effects of language contact. The chapter continues with an account of the spread of Spanish, Portuguese, and French to the Americas, together with the ensuing contacts with indigenous languages and languages of voluntary and involuntary immigration and the formation of Afro-Romance creole languages.