Complex relational units in Danish orthography times especially words from English contribute to complex graphemic units which before 1955 were infrequent or unknown to Danish orthography. So far, the regulation of foreign spellings has concerned mainly words from Greek, Latin and French. CHRISTIAN BECKER-CHRISTENSEN Similar to other orthographies, Danish ”retskrivning” includes complex relational units, graphemic units as <sj> [ɕ] in sjæl ‘soul’, to a lesser extent also phonemic units as [dj] represented by a single grapheme <j> in job. The present paper deals with such complex relations in respect to the delimitation of graphemic units, their function, and their role in orthographic regulation. The investigation relates to the search facility Bogstavlyd (bogstavlyd.ku.dk) and Juul (2010). Three types of complex graphemic units are distinguished: endocentric and exocentric combinations, e.g. <hv> [v] and <sj> [ɕ], and endo-exocentric combinations functioning differently internally and externally in relation to the word as a whole, and the treatment of these combinations as units or sequences in orthographic descriptions with zero correspondences is discussed in the light of minimizing the inventory of graphemic units and providing a basis for general orthographic rules (reference is made to Venezky 1970 and Carney 1994). Some exocentric units are phonographemically necessary units, others are substitutable with single graphemes or other exocentric units. In some more recent loanwords the foreign spelling is functional, marking a sound structure deviant from the Danish main sound structure. Most complex graphemic units contribute to graphemic redundancy and may cause a problem especially to the graphophonemic decoder in recognizing the right delimitation as well as finding the right correspondence. According to Juul (2010), complex graphemic units account for a minor part of sound-letter-sound relations in Danish orthography. Some of the more frequent combinations are units such as <hj, hv, ld, nd, rd> which are not productive in the vocabulary except from compositional and derivational words. On the other hand, in more recent 156 NYS 43 157 Complex relational units in Danish orthography times especially words from English contribute to complex graphemic units which before 1955 were infrequent or unknown to Danish orthography. So far, the regulation of foreign spellings has concerned mainly words from Greek, Latin and French. CHRISTIAN BECKER-CHRISTENSEN Similar to other orthographies, Danish ”retskrivning” includes complex relational units, graphemic units as <sj> [ɕ] in sjæl ‘soul’, to a lesser extent also phonemic units as [dj] represented by a single grapheme <j> in job. The present paper deals with such complex relations in respect to the delimitation of graphemic units, their function, and their role in orthographic regulation. The investigation relates to the search facility Bogstavlyd (bogstavlyd.ku.dk) and Juul (2010). Three types of complex graphemic units are distinguished: endocentric and exocentric combinations, e.g. <hv> [v] and <sj> [ɕ], and endo-exocentric combinations functioning differently internally and externally in relation to the word as a whole, and the treatment of these combinations as units or sequences in orthographic descriptions with zero correspondences is discussed in the light of minimizing the inventory of graphemic units and providing a basis for general orthographic rules (reference is made to Venezky 1970 and Carney 1994). Some exocentric units are phonographemically necessary units, others are substitutable with single graphemes or other exocentric units. In some more recent loanwords the foreign spelling is functional, marking a sound structure deviant from the Danish main sound structure. Most complex graphemic units contribute to graphemic redundancy and may cause a problem especially to the graphophonemic decoder in recognizing the right delimitation as well as finding the right correspondence. According to Juul (2010), complex graphemic units account for a minor part of sound-letter-sound relations in Danish orthography. Some of the more frequent combinations are units such as <hj, hv, ld, nd, rd> which are not productive in the vocabulary except from compositional and derivational words. On the other hand, in more recent 156 NYS 43 157 Filtering and packaging data: The problem of representation Accountability as a general function of language MARTHA KARREBÆK JANN SCHEUER This article discusses some of the challenges involved in textual representation of audio-visual data for discourse and micro-analytic purposes. Today it is more important than ever to pay critical attention to this issue due to the widespread, but often relatively untheorized, use of video data in Conversation Analysis and related fields in the social and human sciences. The article presents a critical discussion of the relationship between audio-visual data, the researcher’s impression of these data, and the textual representation generated through representation practices. The discussion highlights the challenges involved in representation of spoken language and of other types of (non-verbal) information, as well as the complications that emerge when researcher and research participants have different linguistic and/or cultural backgrounds (using research on children as an example). The article argues that representation of research data constitutes an analytic step in the research process, and that it therefore needs serious and critical reflection. In addition, it is argued that the representation is not just an analytic tool but also a way for an intended audience to gain insight into the analytic process. These different functions need to be carefully balanced and considered when transforming audio-visual data into textual representations. The article discusses the concept of accountability. The general assumption is that accountability is a primal concern in language and language use. The point of departure is Harold Garfinkel’s definition in Studies in Ethnomethodology establishing accountability as a key notion in a pragmatist understanding of social life. The term designates the observation that social agents achieve their sociability by acting so that their doings display their own meaningfulness in a social context – thereby creating such a context (Garfinkel 1967). The concept is discussed in three distinctive areas of language study. First, it is demonstrated through an ethnomethodological approach that any form of communication is critically dependent on participants’ mutual recognition of each other’s accountability. Second, some key notions of conversation analysis are demonstrated to be functional in such a perspective. The aim is to demonstrate that the turn-taking mechanics of talk-in-interaction may be conceptualized as methods of constructing accountability. Third, it is demonstrated in a functional linguistic framework how speakers’ accountability is fronted in Danish sentences by such means as word order and the use of a particular paradigm of sentence adverbs. 158 NYS 43 159 Filtering and packaging data: The problem of representation Accountability as a general function of language MARTHA KARREBÆK JANN SCHEUER This article discusses some of the challenges involved in textual representation of audio-visual data for discourse and micro-analytic purposes. Today it is more important than ever to pay critical attention to this issue due to the widespread, but often relatively untheorized, use of video data in Conversation Analysis and related fields in the social and human sciences. The article presents a critical discussion of the relationship between audio-visual data, the researcher’s impression of these data, and the textual representation generated through representation practices. The discussion highlights the challenges involved in representation of spoken language and of other types of (non-verbal) information, as well as the complications that emerge when researcher and research participants have different linguistic and/or cultural backgrounds (using research on children as an example). The article argues that representation of research data constitutes an analytic step in the research process, and that it therefore needs serious and critical reflection. In addition, it is argued that the representation is not just an analytic tool but also a way for an intended audience to gain insight into the analytic process. These different functions need to be carefully balanced and considered when transforming audio-visual data into textual representations. The article discusses the concept of accountability. The general assumption is that accountability is a primal concern in language and language use. The point of departure is Harold Garfinkel’s definition in Studies in Ethnomethodology establishing accountability as a key notion in a pragmatist understanding of social life. The term designates the observation that social agents achieve their sociability by acting so that their doings display their own meaningfulness in a social context – thereby creating such a context (Garfinkel 1967). The concept is discussed in three distinctive areas of language study. First, it is demonstrated through an ethnomethodological approach that any form of communication is critically dependent on participants’ mutual recognition of each other’s accountability. Second, some key notions of conversation analysis are demonstrated to be functional in such a perspective. The aim is to demonstrate that the turn-taking mechanics of talk-in-interaction may be conceptualized as methods of constructing accountability. Third, it is demonstrated in a functional linguistic framework how speakers’ accountability is fronted in Danish sentences by such means as word order and the use of a particular paradigm of sentence adverbs. 158 NYS 43 159 Please say A “Flat a” in radio newscasts 1950-2010 JACOB THØGERSEN During the 20th century, the front allophones of the Danish short /a / and long /a ·/ phonemes are believed to have undergone quite a significant raising in informal language. This study examines raising of /a (·) / in a corpus of radio news readings spanning more than 50 years. In the paper it is shown how raising of /a (·) / has been condemned in prescriptive literature as well as in public debate at least since the late 19th century. Prescription has been particularly hard on the use of raised /a (·) / in the national radio, and here particularly within the news. The paper examines how the articulation of /a (·) / in news readings have developed in the face of this hard prescriptivism. It is found that the front allophones of both /a / and /a · / have raised quite remarkably in the period. In the 1950s front /a / was a low-front vowel; today it is a mid-front vowel. The long front /a · / has undergone a similar raising. The result of these raisings is that /a / today has roughly the place of articulation that /a ·/ had in the 1950s, whereas /a ·/ has moved in on the /a (·) / vowels, the next vowels up. It is concluded that prescriptivism has had a limited effect on these vocalic productions. The paper discusses whether this means that the (previously?) condemned variant has “won”, and concludes that it depends on perspective: Acoustically it has, but this has not led to significant changes in the linguistic system. Sociolinguistically speaking, the raised /a (·) /’s seem to have lost much of their stigmatization, and they seem to be less discussed in public debates, presumably because articulations that were previously stigmatized are today considered to be within the norm. 160 NYS 43 161