Pitch accent

Pitch accent (tone accent) is a feature of certain languages whose variations in pitch (linguistic tones) can be used to differentiate words, but the potentially distinctive tones are restricted to one or two syllables within a word. This contrasts with fully-tonal languages like Standard Chinese in which each syllable can have an independent tone. In a pitch accent language, the syllable with phonemic tone is typically acoustically prominent (stressed). Also, many words may not be marked for tone at all.

Languages that have been described as pitch accent languages include certain Scandinavian and South Slavic languages, Baltic languages, Ancient Greek, Vedic Sanskrit, Japanese, Korean, Yaqui^[1] and Shanghainese.^[2]

It has been claimed that the term "pitch accent" is not coherently defined.^[3] Indeed, it is also used to denote a different feature, namely the use or relatively strong use of pitch, as opposed to other cues like acoustic intensity, to give prominence (accent) to a syllable or mora within a word. The latter feature is dealt with in the article pitch accent (intonation).

Description

Pitch-accented languages may have a more complex accentual system than stress-accented languages. In some cases, they have more than a binary distinction but are less complex than fully tonal languages such as Chinese or Yoruba, which assign a separate tone to each syllable. For example, in Japanese short nouns (1-4 moras) may have a drop in pitch after any one mora but more frequently on none at all so in disyllabic words followed by a particle, there are three-way minimal contrasts such as kaꜜki wa "oyster" vs. kakiꜜ wa "fence" vs. kaki wa "persimmon". Ancient Greek words had high pitch on one of the last four vocalic morae in a word, and since a vowel may have one or two morae, a syllable can be accented in one of four ways (high pitch, rising pitch, falling pitch, none). Also, the mapping between phonemic and phonetic tone may be more involved than the simple one-to-one mapping between stress and dynamic intensity in stress-accented languages.

Proto-Indo-European accent is usually reconstructed as a free^{[nb 1]} pitch-accent system,^{[nb 2]} preserved in Ancient Greek, Vedic, and Proto-Balto-Slavic. The Greek and Indic systems were lost, and Modern Greek has a pitch produced stress accent, and it was lost entirely from Indic by the time of the Prākrits. Balto-Slavic retained Proto-Indo-European pitch accent by reworking it into the opposition of "acute" (rising) and "circumflex" (falling) tone. Following a period of extensive accentual innovations, it yielded a pitch-accent based system that has been retained in modern-day Lithuanian and West South Slavic languages (in some dialects). Some other modern Indo-European languages have pitch accent systems, like Swedish and Norwegian, deriving from a stress-based system they inherited from Old Norse,^{[nb 3]} and Punjabi, which developed tone distinctions that maintained lexical distinctions as consonants were conflated.

Difference

Tone

Firstly, in pitch-accent languages, while the primary indication of word-accent is pitch (tone) rather than stress, there is usually only one tonic syllable or mora in a word, or at least in simple words, and the position of this tone determines the tonal pattern of the whole word.^{[nb 4]} Pitch accent may also be restricted in distribution, being found for example only on one of the last two syllables. This is unlike the situation in typical tone languages, where the tone of each syllable is independent of the other syllables in the word. For example, comparing two-syllable words like [aba] in a pitch-accented language and in a tonal language, both of which make only a binary distinction, the tonal language has four possible patterns:

Tone:

• low-low [àbà],
• high-high [ábá],
• high-low [ábà],
• low-high [àbá].

The pitch-accent language, on the other hand, has only three possibilities:

Pitch accent:

• accented on the first syllable, [ába],
• accented on the second syllable, [abá], or
• no accent [aba].

The combination *[ábá] does not occur.

With longer words, the distinction becomes more apparent: eight distinct tonal trisyllables [ábábá, ábábà, ábàbá, àbábá, ábàbà, àbábà, àbàbá, àbàbà], vs. four distinct pitch-accented trisyllables [ábaba, abába, ababá, ababa].

Stress

Secondly, there may be more than one pitch possible for the tonic syllable. For example, for some languages the pitch may be either high or low. That is, if the accent is on the first syllable, it may be either [ába] or [àba] (or [ábaba] and [àbaba]). In stress-accent systems, on the other hand, there is no variation between different stressed syllables. (If there is secondary stress in a stress-accent language, as is sometimes claimed for English, there must always be a primary stress as well; such languages do not contrast [ˈaba] with primary stress only from [ˌaba] with secondary stress only.) In addition, whereas non-compound words may have more than one stress-accented syllable, as in English, multiple pitch-accent words are not normally found.

Other usage

In a wider and less common sense of the term, "pitch accent" is sometimes also used to describe intonation, such as methods of conveying surprise, changing a statement into a question, or expressing information flow (topic–focus, contrasting), using variations in pitch. A great number of languages use pitch in this way, including English as well as all other major European languages. They are often called intonation languages.

The term "pitch accent" is also used in Native American linguistics to refer to minimal tonal systems such as are found in Iroquoian and Athabaskan languages, for example.

Languages

Ancient Greek

In Ancient Greek, one of the final two or three syllables of a word carries an accent. Each syllable may contain one or two vocalic morae, but only one can be accented; accented morae are pronounced at a higher pitch. In polytonic orthography, accented vowels are marked with the acute accent. Long vowels and diphthongs are thought to be bimoraic, and, where the accent falls on the first mora, they are marked with the circumflex. Long vowels and diphthongs that are accented on the first mora have a high–low (falling) pitch contour, and, if accented on the second mora, they have a low–high (rising) pitch contour.

Examples

English

Hong Kong English

Lexical words in Hong Kong English are assigned at least one H (high) tone. Disyllabic words may have the tone pattern H-o (clóckwise), H-H (sómetímes), o-H (creáte), where "o" stands for tonelessness. Trisyllabic words receive any one of seven possible tone assignments H-H-H (kángároo), H-H-o (hándwríting), H-o-H (róundabóut), H-o-o (thréátening), o-H-H (abóut-túrn), o-H-o (esséntial), o-o-H (recomménd). Toneless syllables receive other pitch assignments depending on their positions: word-initial toneless syllables are M(id)-toned, utterance-final toneless syllables are L(ow), and word-medial toneless syllables vary across two major sub-dialects in the community surfacing as either H or M. Because lexical stipulation of the HKE tones are {H, o} privative, one is easily misled into thinking of HKE as a pitch-accented language. It is probably more accurate to think of HKE as a language with lexical tones.^[4]

South African English

In Broad South African English, /h/ is often deleted, e.g. in word-initial stressed syllables (as in house), but at least as often, it is pronounced even if it seems deleted. The vowel that follows the [ɦ] allophone in the word-initial syllable often carries a low or low rising tone, which, in rapid speech, can be the only trace of the deleted /h/. That creates potentially minimal tonal pairs like oh (neutral [ʌʊ˧] or high falling [ʌʊ˦˥˩]) vs. hoe (low [ʌʊ˨] or low rising [ʌʊ˩˨]).^[5]

Welsh English

A distinctive feature of Welsh English is the rising pitch on the last syllable.

Lithuanian and Latvian

      Baltic
       F  R
      / \/ \
     /  /\  \
    /  /  \  \
   /  /    F  \
  /  /     |\  \
 /  /      | \  \
F  R       L  B  F
Lith.      Latvian

F – falling (acute)
R – rising (circumflex)
L – level
B – broken

Baltic tones are often classified as either "acute" or "circumflex." These labels indicate rather a diachronic correspondence not a phonetic one. For example, Acute is falling in Lithuanian but a high level tone in Latvian and presumably rising in Old Prussian and Classical Greek. Circumflex is rising in Lithuanian but falling in Latvian, Prussian and Classical Greek.^[7]

In the tree diagram on the right, as adopted from Poljakov, names for (original) Baltic tones have been equated with those of modern Standard Lithuanian and the falling tone in Latvian is depicted as derived from a Baltic rising tone. According to some it was Lithuanian that "switched places" the Baltic tones.^[8] This might explain why most languages call a rising tone "acute" while in Baltic terminology a falling tone is "acute." Some controversy surrounds Poljakov's model, it has been harshly criticized by Frederik Kortlandt. Kortlandt contends that broken tone in Latvian and Žemaitian is a reflex of a now disappeared glottal stop in Balto-Slavic not preserved in Aukštaitian (Standard Lithuanian) or Slavic languages and not a recent development of acute.^[8]

Long segments in Lithuanian can take one of two pitches – rising or falling. "Long segments" are defined as either long vowels, diphthongs or a sequence of a vowel followed by a sonorant if they are in a stressed position. Pitch can serve as the only distinguishing characteristic for minimal pairs that are otherwise orthographically identical, e.g., kar̃tų 'time:gen.pl' vs. kártų 'hang:irr.3' (rising and falling tone indicated by a tilde and an acute accent respectively.)^[9]

In Latvian long segments (the same criteria as in Lithuanian) can take on one of three pitches (intonācijas or more specifically zilbes intonācijas) either stiepta ("level"), lauzta ("broken") or krītoša ("falling") indicated by Latvian linguists with a tilde, circumflex or a grave accent respectively^[10] (in IPA, however, the tilde is replaced by a macron because the former is already reserved to denote nasalized vowels.) Some authors note that the level pitch is realized simply as "ultra long" (or overlong.)^[9] Endzelīns (1897) identifies "level diphthongs" as consisting of 3 moras not just two. Broken pitch is, in turn, a falling pitch with superadded glottalization.^[9] And, indeed, the similarity between the Latvian broken pitch and Danish stød has been described by several authors. At least in Danish phonology, stød (unlike Norwegian and Swedish pitch accents) is not considered a pitch accent distinction but, rather, variously described as either glottalization, laryngealization, creaky voice or vocal fry. Some authors point out that the so-called broken pitch is not a pitch accent but a pitch register distinction similar to the ngã register of Northern Vietnamese. See Non-Asian in Register (phonology)

Outside of Central Vidzeme (Standard Latvian) the three-way system has been simplified, in Eastern Latvian (Latgale) only broken and falling pitches are distinguished. Speakers of Rīga Latvian and other more westward varieties differentiate only between level and broken pitches with the falling pitch being merged with the broken one. Thus the Standard Latvian "minimal triplet" or "minimal set" of [zāːle] (hall), [zâːle] (grass) and [zàːles] (medicine) in Rīga Latvian would be reduced to "hall" (level pitch) and "grass" (broken pitch) and "medicine" would be pronounced with a broken pitch just like "grass." Speakers around Ērgļi tend to have just levelled pitch.

In late 19th century Danish linguist Vilhelm Thomsen in the speech of a Livonian sailor identified a characteristic that to him seemed very similar to the Danish stød, this feature was later the subject of research by several Finno-Ugricists.^[11] Although the (Indo-European) Latvian and (Uralic) Livonian are phylogenetically unrelated (being from different language families) both have influenced each other heavily in terms of phonology. Whether Livonian acquired this feature from Latvian or vice versa is debated, however, owing to the fact that Livonian is the only Finno-Ugric language to have this feature, the majority of researchers believe it has been a product of Latvian influence on Livonian and not the other way around.^[11] It is possible that "Livonian stød" would be classified as a pitch accent only by Latvian classification just like the identical Latvian lauztā intonācija, otherwise it would be considered a pitch register, glottalization or similar categories as discussed above.

The Livonian-Estonian-Latvian dictionary at www.murre.ut.ee uses an apostrophe after a vowel to indicate broken pitch, for example, Mi’nnõn u’m vajāg instead of just Minnõn um vajāg.

Norwegian and Swedish

Most dialects differentiate between two kinds of accents. Often referred to as acute and grave accent, they may also be referred to as accent 1 and accent 2 or tone 1 and tone 2. Hundreds of two-syllable word pairs are differentiated only by their use of either grave or acute accent. A list of such Swedish words is collected at Swedish Wikipedia: Lista över svenska ordaccentsskilda minimala par. Accent 1 is, generally speaking, used for words whose second syllable is the definite article, and for words that in Old Norse were monosyllabic.

These are described as tonal word accents by Scandinavian linguists, because there is a set number of tone patterns for polysyllabic words (in this case, two) that is independent of the number of syllables in the word; in more prototypical pitch-accent languages, the number of possible tone patterns is not set but increases in proportion to the number of syllables.

For example, in many East Norwegian dialects, the word "bønder" (farmers) is pronounced using tone 1, while "bønner" (beans or prayers) uses tone 2. Though the difference in spelling occasionally allows the words to be distinguished in written language, in most cases the minimal pairs are written alike. A Swedish example would be the word "tomten," which means "Santa Claus" (or "the house gnome") when pronounced using tone 2, and means "the plot of land," "the yard," or "the garden" when pronounced using tone 1. Thus, the sentence "Är tomten på tomten?" ("Is Santa Claus out in the yard?") uses both pronunciations right next to each other. Another example in Swedish is the word "anden", which means "the duck" when using tone 1 and "the spirit" when using tone 2. "Den heliga anden" could, in writing, be construed as either "the Holy Spirit" or "the holy duck", whereas in speech the pitch accent would convey the intended meaning.

Although most dialects make this distinction, the actual realizations vary and are generally difficult for non-natives to distinguish. In some dialects of Swedish, including those spoken in Finland, this distinction is absent. There are significant variations in the realization of pitch accent between dialects. Thus, in most of western and northern Norway (the so-called high-pitch dialects) accent 1 is falling, while accent 2 is rising in the first syllable and falling in the second syllable or somewhere around the syllable boundary.

The word accents give Norwegian and Swedish a "singing" quality which makes it fairly easy to distinguish them from other languages. In Danish (except for some southern dialects), the pitch accent of Swedish and Norwegian corresponds to the glottalization phenomenon known as stød.

Franconian languages

Extent (orange) of pitch usage in Benelux, Germany and France

A pitch accent is found in the following Franconian languages, or language groups: Luxembourgish, Limburgish, Ripuarian and Moselle Franconian. They are sometimes collectively referred to as West Central German tonal languages. Despite significant differences between individual languages within these groups, generally most often there is a contrast between unaccented syllables and accented ones having either of two tonal accents, or sometimes none. The tonal accents are referred to under a variety of names and denominations. Tonal accent 1, called sleeptoon in Dutch or Schärfung in German, has mostly a steep rise of pitch followed by a sharp drop of both pitch and intensity and a return to their previous levels. Tonal accent 2, named stoottoon in Dutch, is having a linear slightly falling contour keeping intensity level.^[12]

Welsh

Welsh has a simple pitch accent but this does not affect meaning because it is always on the same syllable. It does however make it easier to differentiate words in rapid speech. The stress accent is normally on the last or penultimate syllable in Welsh, but the pitch accent is always placed on the last syllable of a word as a high pitch.

West South Slavic languages

Late Proto-Slavic accentual system was based on the fundamental opposition of short/long circumflex (falling) tone, and the acute (rising) tone, position of the ictus being free as is the state of affairs inherited from Proto-Balto-Slavic. Common Slavic accentual innovations significantly reworked the original system primarily with respect to the position of the ictus (Dybo's law, Illič-Svityč's law, Meillet's law etc.), and further developments yielded some new accents—e.g. the so-called neoacute (Ivšić's law), or the new rising tone in Neoštokavian idioms (the so-called "Neoštokavian retraction"). As opposed to other Slavic dialect subgroups, West South Slavic idioms have largely retained the Proto-Slavic system of free and mobile tonal accent (including the dialect used for basis of codification of modern standard Slovene, as well as Neoštokavian used for the basis of standard varieties of Serbo-Croatian: Bosnian, Croatian and Serbian), though the discrepancy between codified norm and actually spoken speech may significantly vary.^{[13][nb 5]}

Serbo-Croatian

Neoštokavian idiom used for the basis of standard Bosnian, Croatian and Serbian distinguishes four types of pitch accents: short falling ⟨◌̏⟩, short rising ⟨◌̀⟩, long falling ⟨◌̑⟩ and long rising ⟨◌́⟩. The accent is said to be relatively free as it can be manifested in any syllable but the last one. The long accents are realized by pitch change within the long vowel; the short ones are realized by the pitch difference from the subsequent syllable.^[14] Accent alternations are very frequent in inflectional paradigms, both by quality and placement in the word (the so-called "mobile paradigms", which were present in the PIE itself but in Proto-Balto-Slavic have become much more widespread). Different inflected forms of the same lexeme can exhibit all four accents: lònac 'pot' (nominative sg.), lónca (genitive sg.), lȏnci (nominative pl.), lȍnācā (genitive pl.).

Restrictions on the distribution of the accent depend, beside the position of the syllable, also on its quality, as not every kind of accent can be manifested in every syllable.

1. Falling tone generally occurs in monosyllabic words or the first syllable of a word (pȃs 'belt', rȏg 'horn'; bȁba 'old woman', lȃđa 'river ship'; kȕćica 'small house', Kȃrlovac). The only exception to this rule are the interjections, i.e. words uttered in the state of excitement (ahȁ, ohȏ)
2. Rising tone generally occurs in every syllable of a word except the ultimate and never in monosyllabics (vòda 'water', lúka 'harbour'; lìvada 'meadow', lúpānje 'slam'; siròta 'female orphan', počétak 'beginning'; crvotòčina 'wormhole', oslobođénje 'liberation').

Thus, monosyllabics generally have falling tone, whilst polysyllabics generally have falling or rising tone on the first syllable, and rising in all the other syllables but the last one. The tonal opposition rising ~ falling is hence generally only possible in the first accented syllable of polysyllabic words, while the opposition by lengths, long ~ short, is possible even in the non-accented syllable as well as in the post-accented syllable (but not in the pre-accented position).

Proclitics (clitics which latch on to a following word), on the other hand, can "steal" a falling tone (but not a rising tone) from the following mono- or disyllabic word (as seen in the examples /vîdiːm/→/ně‿vidiːm/, /ʒěliːm/→/ne‿ʒěliːm/). This stolen accent is always short, and may end up being either falling or rising on the proclitic. This phenomenon, although obligatory in Neoštokavian idiom and so in all three standard languages, is often lost in spoken idioms from the influence of other dialects (such as in Zagreb, the influence of Kajkavian dialect).^[15]

Slovenian

In Slovenian, there are two concurrent standard accentual systems — the older, tonal, with three "pitch accents", and younger, dynamic (i.e. stress-based), with louder and longer syllables. The stress-based system was introduced because two thirds of Slovenia does not have tonal accent anymore. In practice, however, even the stress-based accentual system is just an abstract ideal and speakers generally retain their own organic idiom even when trying to speak standard Slovenian (e.g. the speakers of urban idioms at the west of Slovenia which don't have distinctive lengths don't introduce that kind of quantitative opposition when speaking the standard language).

Older accentual system, as it was said, is tonal by quality and free (jágoda 'strawberry', malína 'raspberry', gospodár 'master, lord'). There are three kinds of accents: short falling ⟨◌̀⟩, long falling ⟨◌̑⟩ and long rising ⟨◌́⟩. Non-final syllables always have long accents (◌̑ or ◌́), e.g. rakîta 'crustacea', tetíva 'sinew'. Short falling accent can come only in the ultimate (or the only, as is the case in monosyllabics) syllable, e.g. bràt 'brother'. It is only there that three-way opposition among accents is present: deskà 'board' : blagọ̑  'goods, ware' : gospá 'lady'. Accent can be mobile throughout the inflectional paradigm: dȃr — darȗ, góra — gorẹ́ — goràm, bràt — bráta — o brȃtu, kráva — krȃv, vóda — vodọ̑ — na vọ̑do). The distinction is made between open -e- and -o- (either long or short) and closed -ẹ- and -ọ- (always long).

Japanese

Map of Japanese pitch-accent types. Red: Tone plus variable downstep. Green: Variable downstep in accented words. Lavender: Fixed downstep in accented words. Yellow: No distinction.

Japanese is often described as having pitch accent; this differs significantly between dialects. In standard (Tokyo-dialect) Japanese, this "accent" may be characterized as a downstep rather than as pitch accent. The pitch of a word rises until it reaches a downstep, then drops abruptly. In a two-syllable word, this results in a contrast between high–low and low–high; accentless words are also low–high, but the pitch of following enclitics differentiates them.^[16]

Korean

Standard Seoul Korean uses pitch only for prosodic purposes. However, several dialects outside Seoul retain a Middle Korean pitch accent system. In the dialect of North Gyeongsang, in southeastern South Korea, any one syllable may have pitch accent in the form of a high tone, as may the initial two syllables. For example, in trisyllabic words, there are four possible tone patterns:^[17]

Shanghainese

The Shanghai dialect of Wu Chinese is marginally tonal, with characteristics of pitch accent.

Not counting closed syllables (those with a final glottal stop), a Shanghainese word of one syllable may carry one of three tones, high, mid, low. (The tones have a contour in isolation, but for these purposes, it can be ignored.) However, low always occurs after voiced consonants and only there. Thus, the only tonal distinction is after voiceless consonants and in vowel-initial syllables, and there is only a two-way distinction between high and mid. In a polysyllabic word, the tone of the first syllable determines the tone of the entire word. If the first tone is high, following syllables are mid; if mid or low, the second syllable is high, and any following syllables are mid. Thus, a mark for high tone is all that is needed to write tone in Shanghainese:

See also

• Tonal language
• Tone (linguistics)
• Stress (linguistics)

Notes

1. ↑ The term free here refers to the position of the accent, as its position was unpredictable by phonological rules and so could stand on any syllable of a word, regardless of its structure. That is opposed to fixed or bounded accent whose position is determined by factors such as the syllable quantity and/or position such as in Latin where it is on the penultimate syllable if it is "heavy" but antepenultimate otherwise.
2. ↑ Fortson IV (2004:62) "From the available comparative evidence, it is standardly agreed that Proto-Indo-European was a pitch-accent language. There are numerous indications that the accented syllable was higher in pitch than the surrounding syllables. Among the Indo-European daughter languages, a pitch-accent system is found in Vedic Sanskrit, Ancient Greek, the Baltic languages and some South Slavic languages, although none of these preserves the original system intact."
3. ↑ Proto-Germanic had fixed accent on the first syllable of a phonetic word, a state of affairs preserved in oldest attested Germanic languages like Gothic, Old English and Old Norse. Free Proto-Indo-European accent was lost in Germanic rather late, after the operation of Verner's law.
4. ↑ The term is often defined as tone on only one syllable or mora. However, in the Korean pitch-accent system, tone is allowed on two adjacent syllables in initial position, and this contrasts with tone on just one of these syllables.
5. ↑ For example the accentual systems of the spoken idioms of the Croatian capital Zagreb and the city of Rijeka are stress-based and do not make use of distinctive vowel lengths or pitch accent.

References