I have read many descriptions of the lenis/fortis distinction, but they don't make any sense to me.

I have searched in vain for minimal pairs (preferably with audio/vidoe) illustrating the difference between an unvoiced consonant and the "unvoiced" version of its "voiced counterpart".

Does anyone know of such material?

EDIT: The book I'm working with says (my italics):

Even when they are fully devoiced, there is a difference between voiced obstruents and their voiceless counterparts, i.e. [b̥,] is not the same as [p], [z̥,] is not the same as [s], etc. The reason for this is that in German (and English, but not in many other languages, e.g. French) voiceless obstruents are pronounced with greater force and muscular effort than voiced ones. The voiceless obstruents are said to be fortis or strong, and voiced obstruents correspondingly to be lenis or weak. Because of the frequent devoicing of German obstruents, in many situations it is the distinction between fortis and lenis articulation which actually distinguishes pairs of consonants like [s] and [z̥,], [p] and [b̥,], etc., and it seems that for obstruents this distinction is more important than that between voiced and voiceless.

The book then goes on to provide a grand total of zero examples of all of the above.

I have not yet managed to find a single illustration of the claims at the end of the paragraph above, whether recorded or in writing. Such illustrations is what this question is after.

  • 1
    An unvoiced [b] is phonetically the same as a [p]. It's only a /b/ in the sense that the same morpheme may also realize that phoneme as a voiced [b] (e.g. German Urlaub /ˈuːrlaʊb/ [ˈʔuːɐlaʊb̥] == [ˈʔuːɐlaʊp] vs. Urlaube /ˈuːrlaʊbə/ [ˈʔuːɐlaʊbə]). That is, a [b̥] is the same as a [p] but a /b̥/ isn't the same as a /p/, it just sounds the same. Note however that lenis/fortis proposals are about more than voicing/voiced distinctions; often the "fortis" consonant is said to be aspirated, glottalized etc., and then it sounds different (a /p/ that's a [pʰ] sounds different than a /b̥/=[p].) Jul 25, 2018 at 18:40
  • @boiko: please see the EDIT I just added to my post.
    – kjo
    Jul 25, 2018 at 20:06
  • 3
    As I said, when people propose lenis/fortis distinctions, they're generally including features that are more than just voicing/devoicing (aspiration, velarization etc.). For German, consonant/vowel length ratios are important; check out this. Jul 25, 2018 at 20:10
  • @boiko: Thanks for the link. That abstract gave me the very first (and only) example (leid(e)n / leit(e)n) of what this distinction above is all about. I don't understand why it is so hard to come by a decent collection of such minimal pairs. When I search for "german" "lenis" "fortis" minimal pairs none of the hits I've followed contain a single minimal pair, let alone a collection of them, like the one I'm looking for. To get an audio version of the same thing is impossible. If this distinction is as important as the cited excerpt asserts, ... [contd]
    – kjo
    Jul 26, 2018 at 0:32
  • @boiko: ...why is it so difficult to find examples of such minimal pairs?
    – kjo
    Jul 26, 2018 at 0:32

4 Answers 4


My understanding is that in German, the "lenis-fortis" distinction is realized differently in different dialects.

For example, Standard German makes use of aspiration to distinguish syllable-initial /p t k/ from /b d g/.

Some other accents of German do not have aspiration, but may have a contrast in duration.

So the difference between "b̥" and "p" may be something like the difference between [p] vs. [pʰ] in a language like Icelandic or Chinese, or it may be something like the difference beween [p] and [pː].

The difference between "z̥" and "s" may be something like the difference between [s] and [sː].

Interestingly, in some varieties of German that have two series of plosives contrasting in duration but not aspiration, such as Swiss German, the contrast is said to be inaudible in word-initial position when the preceding word ends in a consonant, even though speakers do distinguish between singleton and geminate plosives in production in this context.

Apparently, there may also be a relationship between the lenis/fortis status of a consonant and the duration of a preceding vowel, in the same direction as in English (vowels tend to be shorter before "fortis" obstruents than before "lenis" ones), although it's not clear to me how important this cue is for German-speakers' perception of the "lenis-fortis" consonant contrast. In a paper about "incomplete neutralization" in German, I read that the duration of the vowel in the final syllable of a German word may be different on average before an underlyingly voiceless/fortis and an underlyingly voiced/lenis stop, but in terms of perception, German speakers don't seem to be able to reliably distinguish these kinds of rimes.

Durational differences in the same direction are also supposed to exist for vowels before word-medial obstruents; I don't know if the difference in vowel duration is more consistent or more perceptually salient there.

Minimal pairs

I don't understand why you want to find minimal pairs specifically (standard German obviously has a contrast between "p t k" and "b d g" in some contexts, regardless of how it is realized phonetically, or how important it is for distinguishing words), but if you want them, The German Language Today: A Linguistic Introduction (by Charles Russ, 1994) gives the examples "Paar : bar; Tier : dir; Kasse : Gasse; finden : winden" (p. 117).

  • bad is an interesting example because in the standard pronunciation of the standard language, it is more like bat. In dialect though it is distinguishable. Jul 26, 2018 at 13:57
  • @A.M.Bittlingmayer: In what sense do you mean? I have heard that in some dialects of German vowels were not lengthened before word-final -d, -b, -g, but that seems to be a separate issue. Jul 26, 2018 at 16:32
  • I mean in the standard pronunciation, bad (bathe) is pronounced like bat (requested), Rad (wheel) like Rat (council) and so on. They're homophones. But in dialect they're not homophones. Jul 26, 2018 at 18:07
  • 1
  • @A.M.Bittlingmayer: You're saying they're not homophones because they end in a different consonant sound? I think I've read before that some dialects of German do not neutralize syllable-final obstruents, but I don't remember which areas this is associated with. Jul 26, 2018 at 20:17

The terms "fortis" and "lenis" in linguistics date back at least to the 19th century, and the best you can do is translate the terms as "strong" versus "weak". Consonants are never fortis, or lenis, in an absolute phonetic sense (whereas aspiration, nasalization and so on can be assessed absolutely), instead, this refers to a phonological opposition where one sound is phonetically "stronger" and the other is "weaker". Ejectives are typically termed "fortis", but "fortis" does not entail ejective". Aspirated voiceless consonants are fortis compared to unaspirated voiceless consonants, but unaspirated voiceless consonants are fortis compared to voiced consonants. The terms fortis and lenis are a useful way to deter to a two-way phonological distinction without committing to any phonetic claims. James Foley in his strength-based theory of phonology extensively exploits the concepts of stronger and weaker, devoid of specific phonetic commitments. So one reason why the difference between "fortis" and "lenis" may not make sense is that these are not actually concrete phonetic terms, but we are accustomed to understanding descriptions of sounds in terms of phonetic properties.

The terminological conflict regarding the use of "voiced" in German, specifically, was explained here. The author is equivocating between popular use (where b is voiced=stimmhaftig) and technical phonetic description (where they are unaspirated as opposed to p which is aspirated).

As for the claim about "pronounced with greater force and muscular effort", this cannot be meaningfully demonstrated with any audio or video, except perhaps via an unscientific cartoon. There is no meaningful way of quantifying comparative force or muscular effort behind consonant articulation, and it is not something that can be seen, visually. Nor can you hear a difference in force or effort. In principle you could look at x-ray or ultrasound video of the articulation of different phonemes in some context, but I don't think there are any such videos. The articulatory claim is not based on direct investigation, it is based on an old idea (embodied in the names) that fortis consonants are "stronger" which is widely interpreted as relating to "effort".

  • Thank you for your answer, it contains interesting stuff, but it is also baffling to me, since it completely ignores what my question asked: namely, pointers to examples of minimal pairs illustrating the phenomena under discussion.
    – kjo
    Jul 25, 2018 at 22:42
  • 1
    @kjo I think user6726 was replying to this part of your question: "I have read many descriptions of the lenis/fortis distinction, but they don't make any sense to me."
    – iacobo
    Jul 27, 2018 at 11:49
  • 1
    "t completely ignores what my question asked: namely, pointers to examples of minimal pairs illustrating the phenomena under discussion." What a strange claim. This is one of the best answers you could have gotten. Just take any recording ever of a German sentence and you'll have your so-called "lenes and fortes". It won't help you though, because "fortis" is a concept that evades naming an audible phonetic feature of the speech sound. Dec 14, 2018 at 13:00

This answer is about High German (no dialects).

The standard pronunciation prescribes a clear contrast voiced/voiceless in initial and medial position, neutralisation only occurs at the word final position. It is hard to come up with kind of minimal pairs here, my pick is Wald [vald̥] "forest" vs. walt [valt] (imperative of walten "to rule"). My guess is that German may make a distinction here in extremely careful speech, but not in general. At least, Wald and kalt "cold" are accepted rhyming words despite the lenis/fortis thing involved.

Another is example Waldzwerg [vald̥·tsvɛrg̥] "forest dwarf" vs. Walzwerk [valts·vɛrk] "steel mill, rolling mill", this one is mentioned in handbooks of careful pronunciation for singers, but the main contrast is not fortis/lenis but the juncture in the middle of the word.

EDIT: Additional information (using the 19th printing of Siebs) In reine Hochlautung (the purest form of High German) there are no lenis consonants. The contrast is always voiced & not aspirated vs. voiceless & aspirated. Final stops are completely voiceless. In gemäßigte Hochlautung (relaxed High German) the lenis consonants are allowed in final position and in initial position for speakers with a Southern German accent [b̥] and [g̥] are allowed.

EDIT2: Unlike the creators of the Standard German orthography, the creators of Deutsche Bühnenaussprache were linguistically informed by the standards of their time (late 19th century). At this time there wasn't anything close to a supra-regional standard. Gemäßigte Hochlautung is an attempt to come closer to a Gebrauchsstandard by the current maintiners of Siebs. But it is a kind of chicken-and-egg problem: The Gebrauchsstandard is only there because of Siebs' prescriptionism.

  • 1
    I think Siebs' claims are rather prescriptive than descriptive. The Auslautverhärtung 'final obstruent devoicing' makes me doubt your minimal pairs. Dec 14, 2018 at 12:47
  • I mostly agree. Here's what I wrote above: "fortis" is a concept that evades naming an audible phonetic feature of the speech sound. Dec 14, 2018 at 13:02
  • However, the "standard pronunciation" is usually determined in a descriptive way by looking at a supra-regional standard of usage "Gebrauchsstandard". That is different from prescribing something (To me, the people who do the prescribing seem to be the ones without linguistic knowledge). Dec 14, 2018 at 13:06
  • Okay, I will not deny that. Dec 14, 2018 at 13:33
  • Still, what linguist today still postulates a standard pronunciation that is not based on the supra-regional Gebrauchsstandard? I don't understand how Siebs' claims help us find out what people's pronunciation is actually like. Dec 14, 2018 at 13:35

"Rat" (council), "Rad" (wheel), "Bürger-Rathaus" (village center, mayor's office), "Fahrradhaus" (bicycle house), "Ratschlag" (hint, tip), "Radschlag" (wheel [gymnastics]). The name Conrad supposedly derives from Rath (Con either from coen "king" or from kühn "wise, courageous"; cp. Cohn, cone etc "yewish council, priest"?).

"Rand" (rim), I guess, gave "ranzig" (rancid).

"Bund" (binding [of flowers] union [of countries etc]) and "bunt" (colorful) are so hard to distinguish I'm not sure how to spell "Bun?metall" (wiki says colorful, but since Buntmetalle are ligaments [cf. Legierung] like bronce, abounding in a certain assortment of metalls, there's a mixup I'd say).

"Berg" (hill, mountain), "Werk" (work, product), "Bergwerk" (mine, shaft) but "bergen" (to unearth) "werken" (to work). g/k can be reduced to a glottal stop Ber' and Wer', loosing the distinction. Berg can be lenited to Berch, in one (my own hodgepotch) dialect, but not so for Werk. There's yet a difference between weich (soft) and Weigel (a name) that didn't go to *Weichel but rather *Weiyel (voiced velar fricative [ɣ̞]).

Berliner Ick, Icke seems to retain an old variant over ich (I, ego).

In gucken/kucken the standard couldn't overcome the dialect so kucken is regarded as standard alternative (or alternative standard?).

weg (way, long e) weg (away, short e)--A broad way and a quick away, so to speak--Zweck (purpose) Wecke (bread; southern dialect; surely Weck' somewhere where it's in use, also Weckerl').

Hagen (a name; a kind of place) hager (slim; cp. mager - meager, lean) Haken (hook) haken (to jam) hence poetic hag' and hak'.

mag (I may, like) mögen (to like) vermögen (to be able) mach (I make, do) machen (to make) machte (I did) möchte (i want, would like) -- it almost seems as if möchte was a subjunctive preterite(?) of machen (cp. might, Macht) and mögen a backformation when möchte were reanalyzed as its own verb. Additionally complicated by meiern (to lead, produce, set into motion), mayor and Lt. mega.

"Birken" (birch trees), "wirken" (to effect, work) shows that birch fortified (or vice-versa?). "Henkel" (cup-handel etc.) vs "hängen" (to hang), "Henker" (executer), similarly?

May overlap or clearly differenciated in g/k: Angst (angst, fear) dankst (2. p. sg. to thank) denkst (2. p. sg. to think) langst (2. p. sg. to long) längst (yet; for the longest time). Especially lang (long) may be "lank" in the extreme case.

Tang, Tank shows a clear difference, as do Zange (pliers) and zanke (1. p. sg. to twist, brawl, complain) except in noisy environments, but zankt (3. p. sg.) and *zangt (made up verbalisation of Zange, instead of zwingen (to force, obligate) zwicken (pinch) or whatever) not so clearly. The point is that ng can be stretched, but k not really.

nagt (he gnaws) and nakt (naked) show lengthening of the vowel and subvocal. Nacken (neck) nagen (to gnaw) are fully separated, but I have no clue how to spell nackend [or nakend?] (to be naked, as if participle of a nonexistend verb?). Nagel (nail) and Makel (flaw, blemish) where Makel gave Macke, because Makel invariably sounds shorter than Makel unless specifically stressed as if *Maakel. However, as pointed out in the comments below, Nagel has aspirated (voiced) g.

Genick (neck) and genug (enough) show barely any difference in termination. Neither do gesund (sound and healthy) and Stralsunt (that harbor near Denmark).

Mus (mousse) and Fuß (foot) are specifically differentiated in writing but not audibly so, to me; Muse (muse) /z/ is voiced, Muse (mousses) /s/ isn't. Likewise, Rus (but Russe) Bus (bus, as if *Buss, also Busse) aren't.

Knabbern is colloquial and might, sometimes, as well be transcribed knappern, I suppose (to gnaw, ie. on fingernails), cp. knapp (short), Knabe , Knappe (boy, cnave; difficult etymology).

Topf (pot) may be Topp, short o; toben (to play wild, to storm) 1. p. sg. tobe, tob', long o. That shows the fortition accross the whole word, where a voiced ablauting Toppe became fricative (if I say so myself), cp. Töpfer (potter). Schuppen (shed, shop) may be compared to Schub-lade, Schub (drawer), the later apparently from schieben (push, not pull nor draw) Schub (boost), schubsen (to shove, nudge), but both are compartments so maybe there's a closer relation.

Knabbern is colloquial and might as well be transcribed knappern, I suppose (to gnaw, ie. on fingernails), cp. knapp (short), Knabe , Knappe (boy, cnave; difficult etymology).

Staub (dust) and stopp (stop!) Stab (staff); cp. stopfen (to jam, plug) Stubbel (stub) Stoppel (bristle) Stumpf (stub, trunk) stumpf (blunt) Stube (living room, stove) steif (stiff) Stippe, stippen (dunk, touch) Stepp-Decke (woven blanket) Step-Tanz (step dance). So you can compare your own native stab step stop strobe (if you are native English speaking). Try saying stab-stab-stab quickly like a traicherous villain and stop-stop-stop in return (to reduce the lengthening of the a).

Dieb (thief) Dip, dippen (dip, related to stippen) Tipp, tippen (tip, hint; to tap). You will hardly find minimal pairs because they tended to dissimilate, e.g. tief (deep), long ago.

  • 2
    You've listed countless word pairs like "Rat" (council), "Rad" (wheel). The problem is that they are completely homophonous due to so-called "final obstruent devoicing" (which is not a useful term for German) or "Auslautverhärtung". The term "final obstruent devoicing" is not useful because voicing is usually optional anyway and the final obstruents merge, which subsequently cannot be explained be devoicing. Dec 14, 2018 at 20:41
  • 1
    As far as I understand it, this question is about the following: "Even when they are fully devoiced, there is a difference between voiced obstruents and their voiceless counterparts" Dec 14, 2018 at 20:41
  • You cannot find any minimal pair that differs in homorganic obstruents in final position, because they don't exist in German. The concept that I quoted is about devoiced /b d g/ that are still distinct from /p t k/. Dec 14, 2018 at 20:41
  • 1
    If you think that you can distinguish final /b d g/ in German from /p t k/, then you should conduct a study and publish it, because that would be quite shocking Dec 14, 2018 at 20:41
  • 1
    mainly aspiration and the length of the consonant (compare Jessen, Michael. 1998. Phonetics and phonology of tense and lax obstruents in German. Amsterdam/ Philadelphia) Dec 15, 2018 at 10:07

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.