Spontaneous speech is a cyclic process involving a loosely ordered set of tasks: conceptual preparation, formulation, articulation [LEV 89, REI 00]. Given a goal, we have to decide what to say (conceptualization) and how to say it (formulation), making sure that the chosen elements, words, can be
Chapter written by Michael Zock and Debela Tesfaye Gemechu.
integrated into a coherent whole (sentence frame) and do conform to the grammar rules of the language (syntax, morphology). During vocal delivery, in itself already quite a demanding task, the speaker may decide to initiate the next cycle, namely starting to plan the subsequent ideational fragment. In sum, speaking or acquiring this skill is a daunting task requiring the planning and execution of a number of subtasks. Given some goal, a speaker must plan what to say and how to say it, i.e. (a) find the right words; (b) determine an appropriate sentence frame; (c) put the chosen lemma in the right place; (d) add function words; (e) perform morphological adjustments; (f) articulate.
If speaking is difficult, writing is even a greater challenge, despite the huge amount of extra time. An author must not only know how to carry out most of the operations mentioned, but also be able to perform some additional operations which are not trivial at all. Some of them are at the linguistic level (cohesive devices: links, pronouns, choice of adequate determiner, etc.), and others are at the conceptual level: analysis and synthesis of knowledge, determination of information to provide and ensure reference (Henry Pu Yi, the last emperor, he), grouping, ordering and linking of messages, aggregation, i.e. merging syntactic constituents, etc. These last four operations are fundamental, as otherwise the reader may misunderstand or not understand at all. Being unable to see the connection between the parts, s/he cannot make sense of the whole. The document is perceived as a set of unrelated, i.e. incoherent segments. Yet, texts are characterized by the fact that goals, ideas and expressions are linked via a set of rhetorical (concession, rebuttal, etc.), conceptual (tense, cause-effect, set inclusion, etc.) and linguistic relations (anaphora, reference chains). Indeed, it is quite rare to see “texts” whose elements (propositions or sentences) are related only on the basis of statistical considerations (weight, frequency, etc.).
Text structuring is a particularly challenging task, because the ideas to be conveyed generally lack the links needed to build a coherent topic map, i.e. a tree showing which ideas go together and how the different chunks are related. Moreover, ideas tend to come to our mind in any order, i.e. via association [IYE 09]. Hence, in this case, “order of conceptual fragments” is a by-product of priming. It depends only on the relative associative strength between two items: a prime (doctor) and a probe (target, e.g. nurse). Obviously, this kind of order is very different from the one we see in ordinary texts, where the author guides the reader from some starting point (problem) to the conclusion (end point, solution). In conclusion, the order in which ideas come to our mind is fundamentally different from the order in which they will be conveyed in the final document, a well-structured text. Obviously, this transformation is not an easy task, yet, what makes things worse is the fact that the information needed to impose order on this data is generally absent in the conceptual input, i.e. the messages to be conveyed. This information needs to be inferred. This is probably the reason why writing is so much harder than speaking. Let us illustrate this via some concrete problems.
-  Reading the following sentence: “While there are many similarities between Japan andGermany there are also quite a few differences”, it would be a mistake to believe that theexpressed “facts” are stored like that in our memory. Indeed, what is expressed is probably theresult of an analysis/synthesis of a large set of data concerning these two countries. Once wehave performed this task, we may well conclude that despite the number of commonalities(discipline, work ethic, clean, well organized), there are also quite a few differences betweenthe two countries: geographical location, religion, food (rice/potatoes, fish/meat), behavior(individualism vs. collective behavior), etc.