'L' Deletion

Example of linguistic variation:

•  In spoken French, some speakers 'delete' or leave out 'l' sounds in certain places
• e.g. "Il y a" vs "y a" (often in the case of an impersonal 'il')
• e.g. il va [iva]/ils vont [ivɔ̃]/elle voit [evwɔ]/elles regardent [erəgard]
• /l/ in subject pronouns and in 'word final' positions often deleted
• Content words:
• Possible, Exemple (word final)
• Escalier, Millieu (word median)
• La/Le/Les - word initial position - very unusual for the /l/ deletion to occur here but it does occasionally 
• Lui (e.g. 'je lui parle' [ʒi]0

Historical examples of /l/ deletion:

•  Alterum (Latin) - autre
• Altum - haut
• Alba - aube 
• Bellus - beau (exception - belle)
• Falsus - faux

Grammaticalisation: 

• à + le = au (Old French 'als')
• de + les = des (Old French 'dels')
• jes/quis (Old French) - je les/qui les (Modern French)

Synchronic Perspectives:  

• Ashby - 1988 - Tours
• Armstrong - 1994 - Lorraine Dieuze (metropolitan France study)
• Sankoff & Cedergren - 1976 - Montreal 
• Poplak & Walker - 1986 - Ottawa Hull
• Note: Canadian studies older, Metropole studies more recent
• Studies based on real speech recordings 

Study Results:

•  Geographic:
• /l/ deletion is more advanced in Canada than in Metropolitan France 
• The Metropole is lagging behind in terms of language change
• Metropolitan studies are more recent than the Canadian studies, emphasising the lag
• While the difference is not particularly significant in the extremely frequent contexts (eg il) the gulf becomes greater across the infrequent contexts
• e.g. object /l/ deletion is relatively common in Canada but very uncommon in Metropolitan France 
• Definite article /l/ deletion occurs infrequently in Canada but it has not yet spread to the Metropole
• Linguistic:
•  Lexification of 'il' - the rate of deletion 'il' (personal and impersonal) is reaching 100% (in Canada)  
• therefore a new word has been formed ergo linguistic change has occurred
• The effect which word type has on the frequency of /l/ deletion 
• Pronouns evidence /l/ deletion more commonly than articles
• Within the pronoun category, subjects (il/s, elle/s) demonstrate a greater frequency of /l/ deletion than objects (lui, les, la, le, leur, l')
• Phonological context - deletion was affected by the vowel/consonant pattern
•  e.g. 'il voit' vs 'il entre' - speakers will attempt to preserve the CVCV pattern
• i.e. [ivwɔ] vs [ilɛ̃tr]
• Stylistic - deletion = non prestige 
•  Social Factors (not applicable in Canada - /l/ deletion is universal there irrespective of social factors - another sign of their more advanced stage of language change)
•  Age - always important marker of language change 
• Very high rate of deletion in younger age group, lower rate in older group 
• Armstrong's study of French teenagers showed that the deletion rate was at 100% in certain cases 
• Gender 
• Females delete less frequently than males  
• Males tend to lead the process of language change while women lag behind
• Class
• lower class delete more often than higher class

Conclusion:
Linguistic change is in process. 
  

Liaison as a Sandhi Phenomenon

Liaison - phenomenon of pronunciation, not written
Sandhi phenomenon - change in quality of the sound of a word

Obligatory vs Variable Liaison:

• Les livres vs les‿hommes - follows the French pattern of consonant/vowel/consonant 
• The liaison prevents a double vowel pattern in a sentence
• Liaison not realised before aspirated 'h's eg les halls les haricots, héroes
• Sometimes there's a choice - "c'est‿un village" vs "c'est un village"
• Prescriptive rules - invariable: 
• noun phrase - determiner + noun/pronoun/adj (eg vos‿enfants)
• verb group - pronoun + verb (eg ils‿ont compris), verb+pronoun (eg, allons‿y)
• set phrases - comment‿allez vous, les‿États-Unis etc
• Prescriptive rules - variable: 
•  noun phrase - plural noun + adj/verb (eg des soldats‿anglais, ses plans‿ont réussi)
• verb group - verb + complement (eg, je vais‿essayer, il commencait‿à lire, on est‿obligé etc)
• uninflected words - invariable mono/polysyllabic words (eg, en‿une journée, très‿interessant/pendant‿un jour, toujours‿utile)
• Prescriptive rules - hypercorrect (ie liaison prohibited):
• noun phrase - singular noun + adj/verb (eg, un soldat‿anglais, son plan‿a réussi)
• uninflected words - et + whatever follows (eg et‿on l'a fait)
• special cases - aspirated 'h' (eg des‿héroes), article etc + relevant numerals (eg, cent‿huitième, en‿onze jours, le‿un)
• French speakers have an innate knowledge of when to use the liaison

Segmentation/Processing:

• 'un air' and 'un nerf' are identical sounds (homophones) but native speakers do have ways of distinguishing them 
• 'un nerf' - has a longer consonant to indicate the difference when spoken aloud 
• Children learning French will often attempt to write the liaison - 'un nair' instead of 'un air' 

 Morpho-syntactic link:

•  Question: is liaison use based on morpho-syntactical or lexical rules?
• Seeing as native speakers have innate knowledge of when to use the liaison with any word that falls under the category of liaison possibility, their use of the liaison is an act of generation 
• This suggests that the liaison is based on morpho-syntactical rules

 Types of liaison phonemes: 

•  Phoneme - smallest sound unit to carry meaning
• Most common liaison phonemes:
• nous‿avons /z/
• le grand‿écran /t/
• on‿est /n/
• Rare liaison phonemes: 
• j'ai beaucoup‿aimé /p/
• aller‿en France /r/ (v. rare!)
• du sang‿impur /k/

Enchaînement: 

• Liaison with enchaînement - no hesitation between liaison and following word - i.e. the liaison is 'hooked' directly onto next word e.g. 'tro pambitieux'
• Liaison without enchaînement - hesitation after liaison, e.g. 'trop. ambitieux'

Variation Studies: 

• Average rate of usage - means out of every 100 opportunities to use the liaison, how many of them were realised (NOT what percentage of people used the liaison)
• Ashby - Tours 
• 1981 - average rate of liaison usage - 34%
• 2008 -  average rate of liaison usage - 28% 
• The liaison is evidencing language change, but it is not disappearing - the process of linguistic change is in progress and the end is unclear

Social factors in variation of liaison use: 

• Age - older speakers used the liaison much more frequently than younger speakers 
• Younger speakers only used the liaison at a rate of 11%
• This supports the theory that this feature is changing because younger speakers tend to be quicker to adopt language change, while older speakers are more conservative
• Social Class - middle class speakers used the liaison more frequently than lower class speakers
• classified by occupation into middle and lower class
• Social background conditions us to use liaison in a certain way
• Gender - 1st study - males used liaison more than females
• Unusual because women are generally use prestige forms more often than men 
•  2nd study - pattern reversed - females used it more than males 

 Further comments on Ashby's study: 

• Infrequent liaison use with:
• Forms of avoir,  >0% (e.g. ils avaient. Exception 'ont‿eu')
• Noun + adj, 0% (e.g. 'personnes importantes')
• Verb + complement, rare use (e.g. il regardait un film)
• Pas, 0% (e.g. pas évident)
• Frequent use with:
• Quand - /t/
• Est - /t/
• Negatives - jamais, plus, rien - but not pas 

Prescriptive norms:

• The prescriptive rules for liaison use don't seem to be applied in actual practice
• The study shows that the liaison is used in a very narrow range of morpho-syntactic contexts compared to the full range in which it is possible
• Variation is also sometimes found in contexts considered obligatory by the prescriptive rules
• Failure of prescriptivism

Projet PFC:

• Phonologie due français contemporain - study by Durand and Lyche
• According to their findings, liaison use is frequent in the following contexts:
• Adj + N, e.g. petit avantage
• Adv + Adj, e.g. trop ennuyeux
• Plural N + Adj, e.g. personnes importantes (contrary to Ashby's findings)
• V + Complement, e.g. c'est important