Recent years have witnessed a renewed scholarly interest in the study of writing (as opposed to speech) in a more rigorous and scientific way. In the past decade or so several learned studies have appeared on the subject. Some of these studies have focussed on a comparison of spoken and written language (Gupta, 1971; Greenfield, 1972; O'Donnell, 1974; Poole and Field, 19&6), while in othe studies the focus has been on writing and/or reading (Nystrand, 1981; Rubin, 1979 and Vachek, 1973). Still others have explored the social meaning and cognitive consequences of literacy through a study of writing (Greenfield, 1972; Philips, 1975; Heath, 1980; Scribner and Cole, 1981; and Srivastava and Gupta, 1983).
While the variety and volume of recent work on writing (written language) has been impressive, and several aspects of writing have been subjected to close critical examination, many other crucial aspects and issues have either been merely touched upon, or neglected. Some of these important issues are:
1) The two modes of manifestation (Phonic and Graphic) and their relationship with the linguistic sign;
2) The cognitive and linguistic consequences of the verbalization of the abstract 'langue'/competence in speech or writing; and
3) The decoding of meaning from the written text with or without the mediation of spoken language.
The present paper examines these issues in some detail, and attempts a linguistic study of writing in respect of its form, function, channel and modality.
IIIt was Saussure (1959) who, in the first decade of the present century conceived of the Linguistic sign as being a two-faced entity, uniting the 'signified' (expression). He showed that a linguistic sign involves, simultaneously, a concept (e.g., the concept 'tree') and an arbitrary sound-image, different in each language. This notion was further elaborated by Hjelmslev (1961 : 60) who pointed out "the correctness of conceiving expression and content as coordinate and equal entities in every respect".
However, Saussure had also made it abundantly clear that the units of expression (phonic units) cannot be segmented and delimited without reference to units of content level (conceptual units). According to Saussure "a succession of sounds is linguistic only if it supports an idea, and concepts become linguistics units only when associated with sound-images" (Saussure, 1959 : 3).
The fact that units of expression cannot be segmented except by taking into account units of content (conceptual units) got support from scholars like Hockett, Lamb and Chafe. While relating the grammatical and phonological strata Hockett observed: "To get from the phonological stratum to the grammatical, some king of additional criterion of segregration and identification must be evoked ... in fact, the additional criterion .. is always at bottom semantic, no mater how disguised" (Hockett, 1961 : 46). Similarly, Lamb emphasized the need for referring to conceptual units for the segmentation of the expression level into functional units. According to him: "A language by its nature relates sounds ... to meanings and this relationship is a very complex one which turns out to be analyseable in terms of a series of code like systems, each of which connects two neighbouring strata. The top most structural stratum, the sememic, has units directly related to meaning. These sememes may be thought of as encodable into units of the next lower stratum which, in turn, are themselves encodable, and son on, until one comes out with units directly related to speech or writing" (Lamb, 1962 : 3). Chafe (1962 and 1971) also argues for the position that the phonic units cannot be delimited without reference to the units of the conceptual level.
As is evident from the foregoing statements, Sememes are units which are directly encodable into units related to speech or writing. The controversy rests on the question as to whether the extraction of meaning (decoding) from the written symbols involves the mediation of spoken language, i.e., does language use its graphic resources to represent Meaning-Form relationship directly or via phonology or partly directly and partly via phonology. The literature on the subject, available at present, suggests two separate views on this question. The proponents of the first view hold that reading means decoding of the orthographic symbols first into phonemic representation and then extracting the Form-Meaning relationship. This naturally presupposes a mastery of oral comprehension, i.e., a skill in speaking and understanding. Reading in this context, therefore, becomes an additional skills involving conversion of orthographic symbols into phonemic symbols, and nothing more. This view was most cogently presented by Huey who wrote: "The child comes to his first Reader with his habits of spoken language fairly well formed and these habits grow more deeply set with every year. His meanings inhere in the spoken language and belong but secondarily to the printed (written) symbols" (Huey, 1908 : 1). The view-point can be graphically represented as follows:
Meaning <--> Form <--> Phonology <--> Orthography
The second view holds that both spoken and written entities are two distinct manifestations of the same linguistic unit (form) i.e., they exist for what they represent. Their relationship is not based on the fact that the one merely mirrors the other, but rather on the fact they both are actualizations of the same linguistic unit (Srivastava, 1984). The view that reading requires a via-media of phonology has been rejected by Kolers, according to whom "the questions of interest to the student of reading are not whether all (symbol-sound) correspondences can be characterized by rules, for they can, but whether reading is merely their application. Here the answer is decisively negative" (Lolers, 1970 : 116). This second view may be graphically represented as follows:
It is this second point of view which defines reading as act of decoding of the orthographic symbols directly off the printed (written page (Smith, 1971 and Goodman, 1969).
On the question as to how far orthography represents phonology or vice-versa, there can be two kinds of arguments. The first sees the resources of writing system as merely a device to mirror the structures of phonological units at different levels (a view we have already rejected), and the second maintains that the form associated with meaning as a functional unit is basically oppositional, relational and negative in orientation, and hence free from the constraints imposed upon the material of its realization - phonic or graphic. It is this second view that places writing on equal footing with speech. It further suggests that the common belief amongst a group of scholars that speech is primary, is based merely on the fact of the ontogenetic primacy of speech. It is true that writing is a late cultural development, but once it comes into being, it begins to operate with its own laws and rules. It has been shown by Srivastava and Gupta (1983) that writing as a medium of communication comes at a time in the life-cycle of social institutions when a specific kind of linguistic communication is called for. We all know that for identification in absentia we adduce the evidence of one's signatures rather than one's voice print, and for sending messages across time and space we generally adopt the mode of letter-writing instead of tape-recording (Srivastava and Gupta, 1983 : 536).
Since phonic and graphic material are employed to perform the same function, i.e., externalization of the oppositional and relational aspects of the same form, it is but natural that the two evince certain correspondences. It is because of the correspondences that appear between speech and writing, and because of the ontogenetic primacy of speech, that writing has often been labelled as a derivative of speech. This is a wrong conception because different writing systems externalize the form with different degrees of correspondences between phonology and graphology. For example, the logographic Chinese script externalizes the linguistic form directly, i.e., without a detour via phonology, the phonographic Devanagari and Roman scripts evince an high degree of correspondences with phonology,. Japanese and Korean, which are syllabry-based, come some where between these extreme cases. Homophonous forms being represented differently in writing is an additional proof that writing (despite correspondences of sorts) does not take a detour via spoken language. English has several homophonous forms which are distinctively represented in writing, such as:
Deer - Dear
Knight - NightSimilarly, we have cases where homographic forms are differentiated in speech. Example:
Read - /ri:d/
Read - /red/To conclude this section, we would state that writing as a mode of communication has its own logistics for capturing and externalizing the Meaning-Form relationships that exist in a language as a sign system.
IIIThe differences between spoken and written modes of communication can be discussed in two distinct contexts - communicative message and communicative channel. The message dimension is primarily constrained by the differences in the functions performed by speech and writing respectively (these functions have been discussed at length by Srivastava and Gupta, 1983). It is these functional differences which make spoken discourse highly, context-bound because speech is realised 'within ear-shot' and in a 'locale' defined by speaker-hearer. Written text, on the other hand, cuts across time and space and hence, is highly decontextualized. MOreover, spoken language implies interaction and participatory involvement, while written language evinces integration and detachment. These differences have been worked out in some detail in recent researches (Green field, 1972; O'Donnel, 1974; Poole and Field, 1976; Chafe, 1979; Ochs, 1979 Tannen, 1980, 1982; Olson, 1977 and Rubin, 1978). Based on Ochs (1979) we may summarize the message-oriented differences that exist between planned written text and informal spoken discourse as follows:
K) Greater focus on interpersonal involvement..
In this paper we are more concerned with channel-oriented differences and parallelism between written text and spoken discourse. ON the face of it one finds that cohesion in spoken discourse is achieved through paralinguistic and non-linguistic sign-systems (i.e., voice-quality, tone and intonation, gestures, etc.)., while in written texts it is established by means of different types of syntactic structures and inter-sentence linkers. On the surface representation we find that a spoken word is configurated trough syllable or phoneme-sized units and their combinations, while and punctuation marks.Two things are worth exploring: first, the nature, size and scope of the units that represent the linguistic form and ,second, examining how, on the one hand, phonic material gets structured to vocalize those underlying units and how, on the other hand, graphic material is functionally-oriented to externalize the same units. A still more stimulating area of investigation is to find what aspects of the form remain unrepresented in each mode of representation (phonic and graphic) because fo the constraints imposed by the material of the medium and how a reader/hearer comprehends, even these unrepresented aspects. The third area of interest is the parallelisms that exist between the manifestations of theses two media, e.g., parallelism between prosodic features and punctuation marks.
These distinctions between speech and writing are vital because we find among children a discrepancy between their visual and auditory modalities of learning. Wepman (1968) is of the view that all early learning is modality-bound, i.e., some childre3n suffer from the neurological path ways concerning eye-oriented modality (visiles) and some from neurological pathways concerning auditory processing (audiles). Similar is the finding of Williams (1977) who states:
Some children appear to be eye-oriented and have difficulty with tasks involving auditory analysis. Others seem to be ear-oriented - they have no trouble with auditory tasks, but their visual perceptual abilities are wanting. One approach that has come to be called in the literature "modality matching" suggests that instruction might be more effective for these children if it is geared to their dominant modality (Williams, 1977 : 266-7).
From the foregoing it clearly emerges that both writing and speaking or reading and listening skills must ultimately derive from linguistic knowledge. Speaking and writing are dimensions of performance act, while linguistic knowledge is competence. It is therefore, logical that we connect the performance-act with competence by taking the two kinds of processing - eye-oriented processing and ear-oriented processing - of the same linguistic input. when we try to correlate the two kinds of processing in terms of deep linguistic organization certain interesting differences and parallelisms seems to emerge. These differences and parallelisms become clearer if we discuss them under tow kinds of symbolization:
a) Concerned exclusively with the level of signifier (expression).
b) Concerned with both signifier and signified.
a) Concerning Expression Level
Recent literature suggests that in speech it is the syllable which is functionally coherent unit of expression (Liberman, 1970; Cooper, 1972; Cole and Scot, 1974). Since we articualte syllables and no phonemes, it is appropriate to consider syllable as the minimum unit of production. Acoustic criteria also suggest that analysis and synthesis of syllables is an easier task than is that of segmenting and blending phonemes (Hardy, et al., 1973 and Liberman, et al., 1974). As far as writing is concerned phonemes-sized alphabetic that phoneme segmetation is more difficult for the young child and that this process develops later than syllable segmentation. It is precisely for this reason that syllable-based writing systems (like Japanese 'Kana' - roughly a syllabry) are easier to learn to read than systems based on alphabet (Makita, 1968). The same view is held by Rozin, Poritsky and Stosky (1971).
We therefore, theorize that on the expression level the minimal segmental unit of differentiation is the phoneme and within the phomeme-limit, the phonemic distinction is manifested by distinctive feature. The parallelism between phonic and graphic channels of communication can thus be shown as under:
Expression Level
It is worth mentioning that we find close parallels between the functioning of the units belonging to the two modes of expression. For example, phoneme is defined as a 'bundle of distinctive features'. It should, similarly, be possible to define the grapheme (letter) to be a unit consisting of a bundle of orthographically distinctive features, the only difference being that while distinctive features concerning phonemes are oral-aural in orientation (voicing, aspiration, nasality, etc.), those belonging to graphemes are visual-spatial in orientation. For example, in Devanagari writing system the distinction between the two letters (graphemes) [¯Ö] and [±ú] is marked by the feature ± . That this distinction is functionally recurrent can be shown by taking the visual shape of tow other letters [¾Ö] and [ú]. Similarly, in Devanagari one can posit the distinctive feature [`] which serves to form a minimal pair between [¾Ö] and [²Ö] and also between [¯Ö] and [ÂÖ]. It is worth exploring as to how many such distinctive strokes are there which exhaust all the letters of Devanagari.We get a parallelism between the two sets of units: Phoneme/Allophone and Grapheme/Allopgraph. Scholars have sought to give rules of correspondence between these sets.
[G] <--> /P/
However, since the two systems are two distinct and autonomous realizations, as shown by Gupta (1971) we get generally oblique correspondences as well. For example,
[G] -->/P1, P2, P3 …/ e.g., [C] -->/ s ~ K/ as in cell, cycle, call, come;
[G1 + G2] -->/P/, e.g., [C + h] -->/k/ as in chemist, chrome, etc., or
[G1] -->/f/ e.g., [K] -->/f/ as in know, knob, knee, etc.The search for correspondences between the two autonomous systems and a belief in the primacy of speech, often leads scholars to offer various explanations whenever there is a mismatch between the spoken form and written form of a language. We try to explain asymmetrical relationships such as the ones mentioned above, by saying the writing system is closer- to phonetic reality or, in other cases, closer to phonemic reality. When such explanations are found to be inadequate, we begin to assert that the writing system, in such cases, represents more abstract units of language such as rnorphemic or even lexical reality. For example, Chomsky and Halle (1968) have tried to show that the English spelling system is more abstract in nature and, hence, cannot be explained as a system with simple Grapheme- Phoneme correspondences. It is true that the relationship between Allophone and Phoneme, and that between Allograph and Grapheme is that which obtains between a class and its members, but for establishing the membership of a class, the modes of orientation are different, i.e., oral-aural in one and visual-spatial in the other. For example, the English phoneme /K/ has three allophones - [kh] as in 'key' [q] as in 'cow' and [k] elsewhere. The grapheme [¸ü] of Devanagari has four allographs - [¸ü] as in ¸ü£Ö (rath) [ Ô ü], as in ú´ÖÔ 'karm', [ ? ] as in ÖÏÆü 'grah' and [ ` ] as in ™Òüú 'truck' [¸ü] is considered to be a full letter while the other three are used in conjunct forms. These conjunct forms of [¸ü] as allographs are conditioned by the position the [¸ü] occupies in the compound, on the one hand, and on the specific shapes of the other letters with which it combines, on the other. When initial in a conjunct, i.e., C1 in a C2 C3 sequence, it is written as a semi-circle above the second letter. When it is the second member of a conjunct, i.e., C2 it may be realised as [ ` ] or [ ? ] as shown below:
C2 -->[ ` ] when the proceeding consonantal letter has a rounded bottom as in [™Òü] or [›Òü], and
C2 -->[ ? ] i.e., an oblique stroke to the left of the bar (mid-bar or end-bar) of the proceeding consonantal letter as in [Îú],[±Ïú],[¯ÖÏ],[ÖÏ],etc.Scholars have discussed in much depth the tactic behaviour of phonemes, i.e., phonotactic structure of a given language. These have been discussed under sequential constraints. For example, in English consonant sequences such as /tl/ and /spn/ do not occur initially in a word. In a writing system, how the letters combine with each other, forming conjuncts also forms a set of rules which may be called graphotactic .rules. For example, in Devanagari sequencing of [] relates to the following conditions:
When ["Ö] = C1, either the full form ["Ö] is used with a 'halant', e.g., ["ÖË] or the half form of ["Ö] precedes C2, e.g., ["¾Ö],["ú],["™ü], etc., (the halant form is generally used in type-setting).
When ["Ö] = C2, the full form of ["Ö] is written under C1, e.g., [Œ"Ö], [¯"Ö] (this form is, however, out of vogue now) or the full form of ["Ö] follows half-form of C1, e.g., [Œ"Ö],[¯"Ö] etc.
As regards the syllable, it is one of the most basic units of expression level of language organization. It is worth examining how Devanagari graphology has resolved the problem of syllabry and alphabet. It is interacting to observe that Devanagari is, at -one and the same time, consonantal and syllabic, and on the other hand its primary alphabets are syllable-sized in character. We find in Devanagari the following types of syllable characters:
1) V - type syllable, i.e., vowel alphabet such as [†],[‡],[ˆ ]
2) CV type syllable in which V, the optional vowel - a(schwa), is inherently present as in [¯Ö],[ŸÖ],[ú]= /pa/, /ta/, /ka/. (In case one wishes to delete the inherent vowle 'a' either a diacritic mark called 'halant' is placed below the consonantal character or a secondary form of the same character is used, e.g., [Ëú] or [ú]. A whole set of secondary characters based on the primary consonant characters has been evolved in this manner).
3) Vowels other than the inherent 'a' vowel in CV type syllables takes a composite form of representation with the help of an abbreviated form of that vowel, better known as 'matra'. Because matras as vowel-complements are at least ten in number, a four dimensional visual symbol has been conceived and developed to represent these characters. Thus, for the CV-pattern /ki/ and /ku/ we have the representation [úÖê] and [æú].
The spelling system of Hindi utilising the Devanagari syllabry system represents the deep syllabic reality rather than the one realized at the level of performance. In order to illustrate this point let us take two Hindi words.
1) ×ÃÖ´Ö™ü /sima/ 'shrink'
2) ×ÃÖ´Ö™üÖ /sima/ 'shrank'The orthographic shape of both words reflects the deep syllabic reality by representing three syllables si-, ma and a as [×ÃÖ],[´Ö] and [™ü] for /sima/. Similarly, the three syllables of /sima/, si-, ma- and a are also represented orthographically. In spoken realisation, however, the two words have only two syllables each, i.e., 'si' and 'mat' for 1, and 'sim' and 'a ' for 2.
It is interesting to observe that the prosodic principles of verse formation (chhanda shastra) for all Indo-Aryan languages are based on the deep reality of syllabic existence, i.e., the number of syllables which generate 'VarNik' type of Chhand and the weightage of syllable basically divided into light-syllable (laghu) and heavy syllable (guru). If we look at the syllabic structure of four forms of the lexemes 1 and 2 above we find the following realizations:
Graphic No./Wt. Phonetic No./Wt.
1 1 1a) á ×ÃÖ´Ö™ü ñ si-ma-a 3, 3 1 1 sb) á ×ÃÖ´Ö™üÖ ñ si-ma-a 3, 4 si-mat 2, 3sim-a 2, 4
B) Concerning Levels of Signifier and Signified
Coming to the level involving both the signifier and the signified, we find that the two most important units are 'word' and 'sentence'. It is worthwhile exploring how writing systems, especially Devanagari, symbolise these two units visually. Accoridng to Abercrombie (1965), the object of writing is the 'word'. However, the 'word' as a linguistic unit is generally considered to be a multi-dimensional affair (Chao, 1966 and Mathews, 1974). We have in the literature the concepts of syntactic word ('pada' of Sanskrit grammarians), phonological word and orthographic word, apart from the concepts of 'inflected word, compound word, etc. Phonological word is what is assumed to be uttered between two pauses, while orthographic word is written between two spaces. In Devanagari, the orthographic word is also marked by a top-horizontal bar, e.g., [ÃÖ¾ÖÖ¸ü] , [ÃÖ¾ÖÖ ] , [¾ÖÖ¸ ]. The distinction between two words involving the use of top horizontal bar and spaces can be seen in the following pair:á ¯Öß"Ößü ñ = /pili/ = yellow
ü
á ¯Öß # "Ößü ñ = /pi - li/ = drankThat the two types, i.e., phonological and orthographic word are distinctly operative in a language, and there is a mis-match between them can be seen in the following three instances:
(1) We find certain homophonous words like,
/kha (y)i/ - 'ditch' and 'past perfective
form fo the verb 'kha'.Despite the fact that the two words are identically pronounced, in orthographic representation they retain their separate identity and are represented differently, i.e.,
ÖÖ‡Ô - /kha(y)i/ - 'ditch'
ÖÖµÖß - /kha(y)i/ - 'past perfective form of kha'(2) Certain letters discharge an additional function of discrimination words which are phonologically not differentiated. It should be observed that letters like [ÂÖÍ] [Ö] and [Šú] whcih originally in Sanskrit have the sound values /s/, /¸/ and syllabic / r?/ respectively, are now merged with the / š/ , /n/ and /ri/ respectively in Hindi. Their retention in orthographic words serves the function of a diacritic feature by indicating the Sanskrit origin of the words. For example [ÂÖ™Ëü], though pronounced as /š?a/ parallel to what we get in [¿Öú], i.e., sak 'doubt', marks the first member as being a part of high-flown Sanskrit vocabulary. Similarly, though the symbol [Šú ] is used to represent syllabic mono sound /V/ in Sanskrit, in Hindi it is simply pronounced as CV, i.e., /ri/. While in day-to-day conversation the word [ŠúÖ], i.e., / r?¸ / is pronounced as /rir/, in the writing system it is spelled distinctly from [׸üÖ ] which is the brand name of a detergent.
(3) In certain cases phonologically it is difficult to differentiate between a Noun Phrase and Compound noun, though orthographically they are made distinct - a compound noun is written as one unit with a single horizontal top bar and no space between the two members, i.e., # NN #, whereas an NP is represented as two lexically distinct items, i.e., # N # N # . For example, [üúÖ"Öß×´Ö"ÖÔ ] 'pepper' is a compound noun and hence written as a mono-word, while [üúÖ"Öß ×´Ö"ÖÔ ] 'black chilli' is a noun phrase and represented as two distinct lexical units. Another point to be noted here is that there is no phonological counterpart of the use of hyphen which can differentiate between the two determinative compounds avoiding the risk of ambiguity. For example, [³Öæ-ŸÖŸ¾Ö ] means 'elements of the science of Earth', while [³ÖæŸÖŸ¾Ö] means 'the state of being as element'.Coming to the other unit of representation involving both the signifier and the signified, i.e., the sentence, phonologically it is actualized through intonation and sentential pause, while orthographically it is actualized by a definite mark - a full-stop (.) in the Roman script and a vertical bar 'I' in the Devanagari script. It is often alleged that in speech the rise and fall of pitch, the relative prominence given to a word or phrase, etc., can serve as grammatical signals of completeness or incompleteness and as an index of emotional charge like pleasure, displeasure, surprise, etc. In orthography, the same functions are performed by punctuation marks, though to a limited degree. The delivery of a sentence has two dimensions; one which is central to communication (all those meanings which are communicated by a person speaking in a monotone) and the other, which is peripheral and extra-verbal. The former is represented completely in orthography with its conventions of punctuation, while the latter is represented by certain orthographic devices. For example, a surprised 'Oh' (;) and an emphatic 'no' (italicised or bold-faced) and a guestioning 'yes' (?) can be taken care of by punctuation marks coupled with different orthographic devices such as italicization, bold-face, underscoring and capitalization of letters.
There are certain units which are completely unigue to writing. For example, the concept of tine or paragraph has no slateable correlation with phonological units. According to Crystal, the notion of line is so ambivalent that it calls for experimentation into its functionality. He suggests that 'the line may have a particularly important role to play in interrelating the two main views of reading, the 'synthetic' approach (of letters, words, sentences, etc.), and the 'analytic' approach (of text, paragraphs, sentences, etc.)', (Crystal, 1981 ; 76). Recently researches have been conducted to establish a relationship between the notion of line' and its effect on readability and comprehensibility of a text (Zachrisson, 1965; Fabrizio, et al., 1967; Cromer, 1970; Carver, 1970; Graf and Torrey, 1966; North and Jenkins, 1951). However, the results of these researches have been far from very clear, in that no clear-cut correlation between line and speed of reading or degree of comprehension could be established. It has been vaguely hinted that the line may have something to do with potential of chunking of ideas. As yet no attention has been paid to the linguistic structure that goes with such chunking of ideas.
As far as the mechanics of reading is concerned, the line involves a linear scanning of orthographic symbols from left to right (Devanagari or Roman), from right to left (Perso-Arabic) or from top to bottom (Japanese). This mechanical scanning is coupled with certain linguistic functions, on the basis of which two kinds of lines may be conceived of; end-stopped line in which the last syllable coincides with an important syntactic break, and run-on line wherein there is no matching between the line and grammatical structures. This distinction is most fruitfully and creatively exploited by great masters of literature who ingeniously create the poetic object by establishing varying relationships between grammatical units, rhythmic measures and lines. Let us take, for example, the following two lines:
1 wish a greater knowledge, than f attain
The knowledge of myself: a greater gain
('Christ and Ourselves' by Francis Quarles)Whereas in ordinary prose the run-on line does not require a pause at the end, i.e., t' attain, in verse-form a pause is necessary in order not only to enjoy and appreciate the poem, but also to actualise the rhythmic equivalence of the two
lines and their rhyme-pattern: (i.e., attain: gain). Here the functions of sentence (syntactic unit) and phonological unit (pause expectancy) cut across the notion of line (orthographic unit). Poetic excellence also lies in transforming a run-on line into an end-stopped tine, and in exploiting the interplay between grammatical structures and rhythmic patterns to create the verse-line. Verse lines also perform the function of highlighting certain parallelisms. The Indo-Aryan prosody system is based on the concept of rhythmic measure within the overall patterning of metrical lines (Charan or pad). For example, what is known as Vrtta chhand has four metrical lines with the following sub-categories:1) sama-vr?tta - requires full equipollence, i.e., all four lines should have the same number of syllables.
2) ardha sama-vr?tta - requires semi-equipollence with a matching number of syllables in lines 1-3 and 2-4.
3) viÀam-vr?tta - implies a heterogeniety, i.e., all four lines have different number of syllables.Similarly, the matra chhand of Indo-Aryan prosody is divided into different categories involving the line as a crucial notion for chunking of rhythmic measures into two lines (dvipad), four-lines (catuÀ pad) and more than four-lines (vajra pad).
The notion of line logically leads us to the notion of paragraph which may be defined as a centrally regulated combination of lines displaying structural and thematic relevance. The sentences that constitute a paragraph evince both cohesion and coherence which renders the paragraph a compact: unit with no corresponding phonological counterpart in speech.
To conclude, we find that language and linguistic structures get manifested through two distinct media - aural and visual, and they represent the deep linguistic reality not in an isomorphic manner but in a way restricted and delimited by the form, function, channel and modality that are characteristic of each medium. The notion that says that writing is speaking in print, is too simplistic because writing also involves composing which implicates comprehension by the reader. Speech (talking) is a reciprocal / participatory activity implicating the actual presence of a hearer and his participation in the activity of speaking, while no such participatory presence or reciprocity is essentially involved in writing. In the foregoing we have attempted to relate the orthographic principles to the wider framework of writing and reading.