It is now established that immediate serial recall is consistently impaired when sequences of visual items are accompanied by irrelevant background sound. Immediate recall of sequences of visual items may be impaired if irrelevant auditory stimuli are presented during either encoding or retrieval of the to-be-remembered items. A number of memory tasks, particularly those involving serial recall, are thought to be disrupted by the presence of background sound (Colle & Welsh, 1976, Salame & Baddeley, 1983). Although individuals are aware that the sound is irrelevant to the primary task and are instructed to disregard this input, performance on serial recall is substantially impaired. Typically, irrelevant sounds affect memory for verbal material during serial recall tasks (Jones, 1995). It has been thought that such a paradigm is a really good aim for studies, examining both the immediate memory and the process involved in the involuntary auditory perception (Baddeley, 1996). There is evidence that immediate serial recall of visually presented material is impaired when presentation is accompanied by speech, even when participants are asked to disregard it (Colle & Welsh, 1976, Salame & Baddeley, 1983). On the other hand, when unpatterned white or pink noise accompanies presentation of visual or verbal material for irrelevant serial recall, the effects are not clear. Some studies have reported impairment of recall (Rabbitt, 1968; Wilkinson, 1975), while others found an improvement in performance (Daee & Wilding, 1977; Hockey & Hamilton, 1970).
Irrelevant speech and sound are a common occurrence in many environments, such as driving, crowded offices, dormitory room, hospitals or various kinds of command and control centres (e.g. air traffic control centres). If irrelevant speech impairs performance on cognitive tasks involving primary memory, then job performance may be affected adversely. An understanding of the basic properties of the irrelevant speech effects should facilitate applied efforts.
Firstly, it is assumed that the degree to which the irrelevant material is processed is revealed by extent to interference with. While the effects of white noise on cognitive processing have been generally inconsistent (Jones & Broadbent, 1991), the effects of irrelevant sound on cognitive performance are large and replicable, applying to most individuals. Ellermeier & Zimmer (1997) reported a decline by 30% to 50% of accuracy if irrelevant speech is played during a serial recall task. The reported effects are highly reliable (internal consistency of .85 in Ellermeier & Zimmer, 1997) and also stable (Jones, Macken & Mosdell, 1997).
Since Colle (1980) and Colle & Welsh (1976) proposed what it is known today as the irrelevant speech effect, a large number of experiments have been developed in order to understand its cognitive underpinnings (e.g., Hanley & Broadbent, 1987; Salame & Baddeley, 1982, 1986, 1989, 1990). It has been suggested that only tasks involving memory for serial order would suffer disruption by irrelevant speech (e.g., Jones, Madden & Miles, 1992), yet a different test of this idea presented evidence to the contrary ????(LeCompte, 1994). Most of the work has been concerned mainly with the nature of the unattended sound or its relation to the characteristics of the material held in memory, (e.g., Jones & Macken, 1995d). However, the emphasis of the current study is on how different characteristics of irrelevant background sound affect serial recall memory. Jones, Madden & Miles (1992) proposed a different approach to the irrelevant speech effect. The changing-state hypothesis has been described as the magnitude of disruption caused by an irrelevant auditory background. This background sound is produced by changes in the state from one sound to the next (i.e., acoustic variability) rather than by a continuous sound. The aim of the current study is to establish what are the effects of the changing-state of different songs on serial recall, and also to examine how the meaning of the sound affects these cognitive tasks. This meaning was manipulated through the familiarity condition defined as follows: familiar songs represented meaningful sound for participants, while the unfamiliar ones represented meaningless irrelevant background sound. This can also bring contributions to the ‘interference-by-content’ approach (Beaman, 2004, Neely & LeCompte, 1999).
Jones & Macken (1993) examined the irrelevant speech effect and its implications for phonological coding in working memory. The effect refers to a degradation of serial recall of visually presented items when speech sounds are present. LeCompte, Neely & Wilson (1997) explained that in a typical experiment on the irrelevant speech effect, participants see a series of digits while irrelevant auditory stimuli are presented simultaneously. Recall of these digits is impaired by the presence of the auditory stimuli, even though participants are explicitly asked to ignore what they hear.
The irrelevant sound effect has proved, according to Tremblay et al. (2000), to be reliable and robust. Ellermeier & Zimmer (1997) examined individual differences in objective effects of background sound on performance, in regard to their distribution, temporal stability, and the precision of measurement to be attained. They suggested that individual performance decrements due to noise can be measured in the irrelevant speech paradigm.
Moreover, interference is thought to be one of the most influential constructs in memory research. The ease with which the to-be-remembered items are retrieved from memory is dictated by other stimuli or events that are similar in some way to the target (e.g., Anderson, 2003; Baddeley, 1986, Nairne, 1990, 2002, Neath, 2000). It is possible that the to-be-remembered items are held in working memory as phonological code. Salame & Baddeley (1982) claim that auditory stimuli, including irrelevant sound, automatically enter into a phonological store where they are represented as phonemes, while visual stimuli must be rehearsed subvocally and thereby translated into phonemes to be represented into phonological store. Supporting this, Longoni et al. (1993) also suggested that the phonological loop’s main components are the articulatory rehearsal process and the phonological store. If that is the case, the irrelevant speech could interfere with the articulatory rehearsal process, degrading the information from the phonological store, leading to a poorer performance on trials where irrelevant sound is present. In addition, Baddeley & Salame (1989) proposed a two-component model for the irrelevant speech effect. They suggested that the first component is a filter that passes only speech or sound for further analysis at the post categorical level, and the second component is one where confusion occurs between material of auditory and visual origin on the basis of their phonological similarity.
According to Marsh, Hughes & Jones (2009), distraction by irrelevant background sound of visually-based cognitive tasks illustrates the vulnerability of attentional selectivity across modalities. They aimed to examine whether the principle of interference-by-process can be extended to a setting in which the primary memory tasks involve not serial processing but semantic retrieval. However, they suggested that the mere presence of background sound impaired serial recall considerably. The irrelevant speech effect is thought to result from the seriation of sound sequences, producing competition for the process of ordering the to-be-remembered items. Marsh, Hughes & Jones (2009) also suggested a more functional view of the impairment of retrieval, ‘forgetting’, reflecting the legacy of dynamic and adaptive selective attention processes (such as inhibition: Houghton & Tipper, 1994) that are designed to resolve the conflict during the selection of candidates at retrieval (Anderson, 2003).
Banbury, Macken, Tremblay & Jones (2001) argued that disruption must be determinable to a confluence of processing from the eye and the ear at some level beyond the sensory organs. They described this as a ‘breakdown’ in attentional selectivity. They proposed two explanations for how the breakdown arises: one suggesting that it is an ‘attentional blink’ produced by the sound during which the to-be-remembered items are registered by the brain (Broadbent, 1982). The other explanation of the ‘breakdown’ is made of two subtypes, the first of which proposes that the disruption may be based on a conflict between the to-be-remembered and to-be-ignored items. This could be represented by the similarity in the identity of the irrelevant sound to the items being retrieved (Salame & Baddeley, 1982). The other subtype is known as the changing state hypothesis, a part of a model of working memory – the object-oriented episodic record model (Jones, 1993, 1999). The disruption is thought to be greater when a sequence of changing sounds is present, rather than repeated irrelevant sounds. Kllate et al. (1995) also suggested that music containing changes in tempo or pitch produces higher impairment in cognitive performance, compared to music with legato passages.
Although the changing state effect is empirically robust, there are studies that do not support Jones’ (1993) hypothesis. Generally, it was shown that the greater the change in sounds, the greater the disruption, but there may be an upper limit to the changing state. It was suggested by Jones et al. (1999) that if the degree of change is very big, the degree of disruption can also decrease. In addition to this, Banbury, Macken, Tremblay & Jones (2001) claim that sequences of sounds that are composed of different notes from different musical instruments, where the acoustic variation is great, can produce less disruption that does a sequence of sounds where the differences are smaller and the different notes are made up of the same instrument.
Jones et al. (1992), when examining the serial recall as exposed to changing state sounds, manipulated the sound such as if this was a stream composed of a single differing syllable drawn from a small set. Their results concluded that only the varying sequence produced considerable disruption in serial recall. This suggests that there is a significant effect of impairment of serial recall only by the irrelevant background sound that has ‘high changing state’ known as a varying sequence.
As evidence supports that meaning does have a significant effect on serial recall, the current study manipulated the meaning of the irrelevant background sound through familiarity. although participants were asked to disregard the background sound, the prediction is that familiar songs cause greater disruption on serial recall than unfamiliar songs do, as participants’ selective attention may be affected when participants hear a familiar song.
It has been argued whether meaning of irrelevant background sound affects serial recall memory. Oswald, Tremblay & Jones (2000) investigated whether the presence of meaning increases disruption of performance in cognitive tasks. They suggested that both meaningful and meaningless sounds disrupt the cognitive task, but the effect of meaningful speech was significantly greater. There is also evidence that when meaning of the irrelevant sound is manipulated, the disruption of serial recall increases marginally (LeCompte, Neely & Wilson, 1997), but also that there is no impact of meaning on cognitive tasks (Buchner, Irmen & Erfelder, 1996; LeCompte & Chaibe, 1997). LeCompte, Neely & Wilson (1997) conducted a study examining the importance of speech to the irrelevant speech effect. They provided evidence that meaningful speech disrupts recall more than does meaningless speech.
It was suggested that irrelevant speech effect is a mere consequence of auditory stimuli that have access to the same space as the to-be-remembered items (e.g., phonological store – Burgess & Hitch, 1992; Gathercole & Baddeley, 1993; Salame & Baddeley, 1982; primary memory – Neath, 2000). Although these models differ in how interference arises, they all provide evidence for the interference-by-content approach, suggesting that recall is impaired as a result of similarity in identity between the to-be-remembered and to-be-ignored items.
On the contrary, there is some evidence showing that the disruption is caused by process, not content. Marsh, Huges & Jones (2009) examined the interference-by-process approach, suggesting that it determines the semantic auditory distraction, rather than interference-by-content. They argue that interference-by-content approach fails to acknowledge the importance of the nature of the primary task. The impairment of recall is mainly determined by the co- existence of similar to-be-recalled and to-be-ignored items within the phonological store.