Week 11 Flashcards
Hierarchy relations between events, or adjacent events, structure of sequence, first level representation which is
Syntax
Hierarchy atemporal
Or hire chary as an event structure
IMp.
Events based on importance, stored in long term memory as a tool kit. As abstract knowledge structures. Key hansel, atemporal
Event structures are temporal. Tree diagrams.
Musical syntax
Language
The first question is much easier to answer. Most of the world’s music is syntactic
with identifiably discrete elements and norms for the combination of these elements.The last two questions are a lot more difficult to answer and will be the main focus of today’s lecture. In the past, there has been skepticism regarding the lack of
correspondence between music and language, particularly the lack of linguistic parts
of speech such as nouns and verbs in music; however, more recent research indicates important connections between linguistic and musical syntax.
Structural richness of language
Multiple layers, hierarchical, recursive,
Relationship between syntax and meaning
Meaning related to order.
Content dependant structural functions
Ani Patel (2008) states “Linguistic syntax is remarkable for its structural richness,
attaining a level of complexity that sets it apart from any known nonhuman
communication system.”
He puts forward 3 important aspects of the richness in language.
1)
2)
3)
Multi-layered organization involving hierarchical and recursive structures.
Strong relationship between syntax and meaning, so that changes in the order ofwords greatly alter the meaning of a sentence.
Words take on abstract grammatical functions that are determined by their
context rather than inherent properties of words themselves
Multi layered organization in music
Ani Patel states “One of the principal features of linguistic syntax is that relationshipsbetween words are not simply based on nearest neighbour relations.” He presents
the sentence “the girl who kissed the boy opened the door.” An English speaker knows that the boy did not open the door even though the last words are “the boy
opened the door.” Words are not interpreted in a simple left-to-right fashion, but through combination of words into phrases in a hierarchical nature. Syntactic tree
diagrams (different than prolongational tree reductions) display this organization.
Western tonal music deals with hierarchically organized units and subunits that may be temporally distant. One example of a hierarchy within abstract knowledge
structures is the tonal hierarchy. Krumhansl’s theory defines the psychological relations among tones, chords and keys in a hierarchical mental schema. An event
structure hierarchy is therefore based on the tonal hierarchy.
Hierarchy or surface processing more important, Bigand
Tension vs relaxation
Fragmence condition. 12, q123,1234 Chronologies l'ordre 12, 23, 34 Random Random order with transposition, cannot build higher representation Stop tension task
An experiment by Bigand & Parncutt attempted to predict the perceived musical tension in long chord sequences by hierarchical (tonal hierarchy and Lerdahl and Jackendoff GTTM) and sequential models (chord-to-chord). In the “fragments” condition, chords were played in order, and the sequence was stopped at a given chord. Listeners indicated the degree of musical tension they perceived, which is considered the inverse of musical stability. Then the sequence is repeated adding another chord at the end, and so on (e.g., 1-2, then 1-3, then 1-4, then 1-5, etc).
There were 3 other conditions to emphasize the distinction between local and globalharmonic structure. In these conditions, pairs of adjacent chords are played and the
musical tension is rated. In the “chronological order” condition, the pairs are played in the same order as in the piece (e.g., 1-2, then 2-3, then 3-4, etc).
In the “random order” condition, the pairs of chords are played in random order, thusperturbing the normal harmonic context (e.g., 9-10, then 27-28, then 3-4, etc). The
“random transposition” condition was identical to the chronological order condition
except that each pair was randomly transposed to a different key. The average
tension values obtained for each chord are compared across the four conditions.
The hierarchical model predicted that the degree of musical tension associated with a given chord reflects the hierarchical importance (stability) in the prolongational tree.
Shorter branches terminate on less important pitches, whereas longer branches terminate on more important and structural pitches. For example, the second chord (right-branch) would be perceived as subordinate to the first chord and therefore more tense. The tension values for the model were applied to a given chord based on hierarchic distances to the previous chord and in relation to the first chord. Therefore, the section outside of the home key (marked “other key excursions” above) would be predicted to be perceived as higher in tension due to the distance from the home key. Notice the “other key excursions” branch off from chord 17 (in the home key) and do
not extend as far in the hierarchy.
Only a small effect of presentation order.
Highly all correlated, surface more imp. Suggested.
Bigand findings?
These are the average tension values perceived by musicians (top) and non-musicians(bottom). The effect of global structure was weaker than expected, and the average
ratings observed in the fragments condition correlated with the other conditions
(that were manipulated to highlight the local context).
Purple boxes indicate the chords in the home key. It was predicted that the area outside of the home key would be perceived as more tense compared to the home
key sections; however, the return to the home key was perceived as a departure from the local tonic, generating tension (see chord 32) rather than the expected effect of
relaxation.
Strong relaxation occurred when two adjacent chords formed a perfect cadence (V-I), as in chords 2-3, 16-17, 30-31, 32-33, 34-35. Tension was at a medium level on
dominant chords and was high for all other kinds of harmonic progressions, suggesting that listeners were sensitive to the presence of harmonic cadences at the
local level.
Overall, Bigand and Parncutt found that the tension profiles of their listeners were well modeled using local harmonic structure (especially cadences), with a negligible
contribution of hierarchical structure. There was only a small effect of presentation order, which may indicate that the global context did not play a strong role when
making explicit judgments in the “stop-tension” task (an unnatural listening situation). The methodology may have encouraged moment-to-moment listening
within a small sliding window.
Cadences most imp. Stable ending, low tension.
Bottom up vs too down processes
Forced listeners to hear in a small window, forced to think ending in the Bigand study.
Continuously rate tension on slide.
Modeling tonal tension. Changing task hanged conclusions. Bother parameters play equal role.
The following examples show the importance of top-down processes or knowledge
driven processes that are context dependent.
Top-down and bottom-up processes are indispensable for a complete understanding of the environment. Sensory driven processes ensure the cognitive system is informed about signals of the environment while knowledge-driven processes facilitate the processing of both simple and complex signals.
In Western music, syntax-like rules are rooted in the psyc
State someing weirded about the Bigand task, continuous task needed!
Context dependant structural functions
Specific to language inherent properties of words not important, like timbre not imp. . Tonal languages though do care about it.
Tonal language autism I wonder?
Music is focused on inherent properties, and on context.
Relations with overall context most important.
Something that eveloved over time, at one time sensory part was more important
Context dependant strutrql functions in muisic
In this study by Bigand, notes 1-23 are identical in the two melodies with the exception of notes 1 and 20 (circled). The context established by the opening bar determines the tonal context within which the tonal function of the notes is interpreted, and their functions determine the degrees of stability and tension they are perceived to carry when the melody stops on that note (e.g., stop note 2 is the
stable tonic in A minor but less stable scale degree 2 in G major). The underlined
notes indicate the notes of the tonic triad in each key. The A melody concludes on astructurally important note (tonic) while the G major melody concludes on an
unstable note (scale degree 2). The stability profile was predicted to vary strongly
between the two melodies even though the melodies share most pitches and have
the same contour and rhythmic profile. Participants rated the degree of stability on each stop note starting from the beginning and moving forward, similar to the
“fragments” condition of the Bigand and Parncutt (1999) study.
Legend: Melody T1 (A minor)) and Melody T2 (G major).
The purple circles indicate the notes that were actually different between the two
versions of the melody (G# in A minor and G in G major).
The stability profiles are negatively correlated, which shows that listeners perceived
the pitch structure of the two melodies differently, despite having almost identical
pitch structure and identical contours and rhythms. The results underline the strength of cognitive top-down processes on the perception of melodies.
Event structure processing
Relations in old an interpretation onfuntion with respect to activated abstract knowledge stru truces which in turn body mental schemata that anticipate possible continuations ANticipatory schemata Generate expectations Orient perception Facilitate perception
Determine stability of events in large structured
Musical context very primed
Priming paradigm
Two levels of expectation, high or low, other study found middle,
Fast accurate, simple perceptual task.
The priming paradigm is an attempt to get away from explicit ratings by listeners. A context is presented, then a target event (a chord in this case), and then listeners have to make a simple judgment about the target chord as quickly as possible: is it in tune or out of tune? When it is out of tune, one of the notes is mistuned by a quarter-tone, making it very dissonant. So the task is very simple. The question is whether the preceding context will affect performance on this simple task in terms ofreaction time and accuracy of judgment (i.e., getting it right or wrong).
When the upper context is presented, the chords should be interpreted as a V-I cadence. When the lower context is played, the same chords have a harmonic function of I-IV, which is a less expected ending in classical music. According to this paradigm, one would expect faster, more accurate responses to the in-tune/out-of-tune question in the upper context than in the lower context. And indeed this is whatBigand and Pineau (1997) found. This suggests that the processing of harmonic spectra is facilitated for events that are most predictable in the current context.
But a big question of particular relevance to our interest in the effect of context on the interpretation of musical function is whether contextual information presented before the 6-chord sequence (more global relations) can change the expectedness of the two endings and thus reactions to the target chord.
Bigand and colleagues went even further. They added another set of chords in front
of the first set, thus extending the context, still giving V-I or I-IV interpretations (top
and middle rows, respectively). They then created a hybrid that had the immediatelypreceding context from the weakly related sequence and the initial context from the highly related sequence. They predicted that this context would give a degree of
expectedness somewhere in the middle. This facilitation would indicate that the processing of a target chord had been primed in this third sequence by the very beginning of the sequence (the first chords are highly related).
In line with the predictions, highly expected chords were reacted to faster and more accurately than unexpected chords, with the chords having a medium degree of
expectation in the middle. Nonmusicians were slower and less accurate than
musicians, but the pattern of results is identical. These results show that mental
schemas generated by the context affect the processing of the chords and that more expected events are processed more accurately and more quickly than unexpected
events. The syntactic function assigned to the chords depends on this context. Thesefindings further suggest that context effects can occur over longer time spans and at several hierarchical levels of the musical structure.
