June 06, 2005

Blog or Block?

Has my joke been spoiled by an inconvenient fact? I recently observed that the French have been instructed by the authorities to pronounce "blog" as if they were speaking German. But Lucas Champollion wrote in to explain that

Actually, we don't devoice "blog" - otherwise it would sound like "Block", which has the same meaning as its English counterpart. Even if the context might be enough to disambiguate, it just wouldn't sound serious.

Lucas admitted that his judgments might be affected by the fact that he has been speaking mostly English for a year or so, so he provided a scholarly citation (Caroline Féry, "Final Devoicing and the Stratification of the Lexicon in German", HILP 4, 1999), which mentions "rave" (p. 14) as another example of a recent foreign borrowing that resists final devoicing.

This surprises me a bit, since I had always thought that German final devoicing was a phonologically transparent process, somewhat like the process of flapping and voicing in American English that makes "latter" sound the same as "ladder". Among other things, it's not uncommon to hear a German native speaker devoicing at least some final consonants while speaking English. This is a stock feature of caricatured German accents ("Ve haff vayss to make you say 'bloc'...").

So now I wonder what's really going on in German. Is final devoicing variable in borrowed words? Is it perhaps also a gradient process, so that specific instances might be partially but not completely devoiced? I welcome observations from German speakers (you can send email to myl@cis.upenn.edu), but references to empirical studies would be even more welcome. Judgments about variable and gradient phenomena of this type are rather unreliable. For example, I believed for many years that my pronunciations of word like "latter" and "ladder" were distinct, until I discovered by trying the experiment that I'm unable to distinguish my own normal productions at better than chance level.

[Update: Joe Salmons emailed:

German final devoicing (or fortition in a growing number of recent views, like those of Iverson & Salmons, Jessen and others) is almost certainly a complete neutralization (Fourakis & Iverson, Jessen, see now Piroth & Junker) when speakers are unaware of the task in an experiment. F&I argued (persuasively, I think) that speakers who are aware of the task, can produce differences here -- they have not only orthography but also alternations involving these pairs and most German speakers have some knowledge of a language with final laryngeal distinctions.

On loan words, there's long been controversy -- brav and naiv are the classic examples, like rave with a final /v/. My sense has long been that this is a matter of integration of loans: Most speakers have categorical fortition and do it here too. Some speakers, though, produce this as a basically French- or English-like rendition. The former pattern is more widespread, but my impression is that the latter is spreading, esp. among those who are very proficient English speakers. I don't have acoustic data, but I vaguely recall that it's been argued (maybe just in conversation somewhere) that where they do this, it's mostly vowel length that distinguishes the pairs (versus glottal pulsing, for example).

And speakers most definitely carry this over into English. Purnell et al. (along with other works in the pipeline) look at apparent devoicing/fortition effects in Wisconsin English and our German-speaking control showed complete neutralization in English. This is somebody who's lived in the US for ca. 40 years and speaks very good English with a clear accent. I'm pretty sure that the perceptual results for his productions showed nothing beyond chance at all, but Tom will correct me if that's wrong.



Fourakis, Marios, and Gregory K. Iverson. 1984. “On the ‘Incomplete Neutralization’ of German Final Obstruents.” Phonetica 41:140-49.

Iverson, Gregory K. 1997. Review of Final Devoicing in the Phonology of German, by Wiebke Brockhaus. American Journal of Germanic Languages and Literatures 9:255-64.

Iverson, Gregory K., and Joseph C. Salmons. 1995. “Aspiration and Laryngeal Representation in Germanic.” Phonology 12:369-96.

———. 1999. “Glottal Spreading Bias in Germanic.” Linguistische Berichte 178:135-51.

———. 2003. “Laryngeal Enhancement in Early Germanic.” Phonology 20:43-74.

Jessen, Michael. 1998. Phonetics and Phonology of Tense and Lax Obstruents in German. Amsterdam: Benjamins.

———. 2001. “Phonetic Implementation of the Distinctive Auditory Features [voice] and [tense] in Stop Consonants.” In Distinctive Feature Theory, ed. Tracy Alan Hall, 237-94. Berlin: Mouton de Gruyter.

Jessen, Michael, and Catherine Ringen. 2002. “Laryngeal Features in German.” Phonology 19:189-218.

Piroth, Hans Georg, & Peter M. Janker. 2004. “Speaker-dependent Differences in Voicing and Devoicing of German Obstruents.” Journal of Phonetics 32:81-109.

Purnell, Thomas, Dilara Tepeli & Joseph Salmons. 2005. German substrate effects in Wisconsin English: Evidence for final fortition. American Speech 80.135-164.

And Geoff Nathan wrote:

I certainly remember Ulli Dressler, while talking about the difference between natural processes and rules (several hundred years ago at the LSA Institute in Maryland), attempting to pronounce final obstruents and literally being unable to do so without the addition of an audible and fairly forceful final schwa.  'See', he said, 'I just can't say [bundə];  I mean [bunt].  No, [bundə]'   And it wasn't as if Ulli couldn't give us a disquisition on the physiology, phonology and whatever of the process involved....

It may be that younger speakers, under the influence of lots of English are losing that constraint (we Natural Phonologists used to say 'process'), but it sure felt pretty categorical twenty years ago.

This strengthens my belief that the pattern Lucas reports, while perhaps found in younger, more English-aware speakers, is not typical of traditional German pronunciation norms. ]

[Update #2 -- Philip Newton wrote:

I'm a native speaker of both German and English and was born and grew up in Germany.

In response to http://itre.cis.upenn.edu/~myl/languagelog/archives/002228.html , I'd like to say that in my experience, German final devoicing is ubiquitous, even when this causes ambiguity.

For example, in my work as a software developer, there is occasional confusion between "Logfile" (log file, file containing diagnostic output) and "Lockfile" (lockfile, file used to co-ordinate several processes sharing one resource), since they are pronounced identically, even though differentiating them would be beneficial.

This sometimes leads to misspellings such as "Life-Übertragung" for "Live-Übertragung" (a television transmission that is broadcast live).

I can't think of a borrowing with a voiced final consonant. (I could image, though, that they might exist, especially in words coined by youths who are comfortable with English.)

I don't know of any cases where a phonological process (such as German final devoicing, or English flapping) is suspended in ordinary speech just in order to avoid ambiguity. There is certainly such a thing as "facultative pronunciation", where a speaker may use special means (including cancelling some normal assimilations, lenitions, fortitions etc.) in order to convey a distinction that would otherwise be unclear. But that's different. ]

Posted by Mark Liberman at June 6, 2005 06:38 AM