The limitsations of short-term memory: How does memory affect our writing of numbers?
Read Hadar’s full dissertation (in Hebrew)
Short-term memory (STM) allows maintaining information for short periods of time. This ability is critical when comprehending numbers and writing them: in many such tasks, e.g., writing down a telephone number, copying numbers, or writing numbers to dictation, we must maintain the number in STM until we write it or use it. Indeed, several studies have shown high correlations between STM and number writing skills. Nevertheless, the precise nature of this relation is still poorly understood. Here, we examined the role of STM in number comprehension and dictation: we showed the existence of a phonological input buffer for number words – an STM process in the verbal-phonological input stage, we showed that this process is necessary for good number processing, and we showed how it works.
We report two individuals with dysnumeria – a disorder in processing symbolic numbers, who made many errors in number dictation. To detect the origin of their errors, we first examined their performance in several tasks that tap different processing stages. Both participants showed poor performance in tasks that tap verbal-phonological input, but good performance in tasks that tap production of digit strings and the core representation of numbers. We conclude that their deficit is at the verbal-phonological input stage. Second, we showed they made both lexical errors (digit substitutions) and syntactic errors (disruptions of the number’s structure). We conclude that their deficit is in a process that handles both lexical and syntactic information, such as the phonological input buffer. Third, we showed that their deficit is manifested mainly in tasks that create STM load, indicating that their deficit is in an STM process. Overall, these results show that the participants’ deficit was in an STM process at the verbal-phonological input stage, which maintains both lexical and syntactic information – in other words, a phonological input buffer. The results therefore demonstrate the existence of this buffer.
Based on a combination of these conclusions with findings from previous studies, we propose a detailed cognitive model of the processes involved in the verbal-phonological input of multi-digit numbers. The model describes in detail the role of the phonological input buffer, as well as several lexical and syntactic processes at this stage.
This model, and the tools and methods developed in the present study, can help diagnosing specific types of dysnumeria. Precise identification of buffer-related deficits might also open the door to the creation of educational intervention strategies to improve STM and number writing skills. The study could therefore benefit educational environments, by affecting teaching methods aiming to improve numerical literacy and memory.