1 - Scientific background and objectives
Researches conducted in reasoning and problem-solving areas over the last decades both point to the main idea that the content of problem influences its processing: Human being bring their own knowledge into every situations, and used them to comprehend those situations, to reason on them and to achieve their goals. Understanding of the implications and applications of this phenomenon is central to increase our understanding of knowledge acquisition process. Indeed, one of the major purposes of teaching is to give to pupils the capacity to conceptualize situations; that is, to give them
“tools” to reason independently of the context and go beyond the spontaneous semantic coding of the situation.
This project deals with the area of basic school learning. It aims at examine, from a comparative point of view, the influence of content effects on arithmetic problem-solving, which is considered as a foundation for school knowledge acquisition, and to apply the results obtained to promote acquisition.
This proposal opens up new perspectives for teaching since it is designed to show that a major
component of learning is the capacity to go beyond a spontaneous interpretation, driven by semantic dimensions implied by the problem content, to an adequate interpretation that is compatible with the formal structure of the problem. Methods that are proposed are designed to promote a change of point of view regarding the situation.

2- Description of the project, methodology
The originality of the methodology used comes from the experimental measures that will be taken into account. In addition to the common success/failure rates, strategies used as well errors made by pupils will be interpreted as cues that permit to identify the underlying interpretation. Experiments will be carried out with two types of population. Elementary school pupils from both suburban and priority teaching areas will be tested and this from first grade to fifth grade pupils. In
order to highlight elements relative to potential explanation of school failures as a function of type of problems, analyses will be conducted with a comparative point of view. The entire set of experiments breaks down into two parts.
In the first part, content effects on different types of problem will be studied and will lead to the elaboration of a data base that will be made available to scientists and educational persons. A set of semantic dimensions, for which an effect already has been shown in other contexts, will be manipulated and their influence on problem difficulty, strategies used and errors made will be measured. The data base will be constituted from the data collected in these studies and access to it will be given by a web application. This data base will allow the results to be consulted at any time and will be updated as new data will be available.
The goal of the second part of this project is to translate into teaching purposes the effect of semantic content of the problems. Semantic dimensions of problems will be manipulated in order to facilitate the construction of an interpretation of the problem compatible with its formal structure. Series of experiments will be carried out to assess the feasibility of this approach. These studies have direct school implications, especially for pupils with learning difficulties, but also in the area of Intelligent Tutoring System (ITS).
The ITS perspective is also a main component of this project. The ITS that will be developed will allow to obtain a fine diagnosis of each pupil production that will also be made available for the teacher. Furthermore, the remediation component based on the experimental results from both parts of the project will present to the pupil learning situations in accordance with the automatic diagnosis;
thus, it intents to promote the development of learners’ transfer capacity.

3- Expected results
The following results are expected:

• Systematic and well controlled experimental results on a large group of subjects that should highlight the influence of content effects in arithmetic problem solving;
• Experimental results relative to a learning method which takes into account content effects;
  
• Elaboration of a problem data base made accessible by a website that will allow to measure content effects
• Achievement of a cognitive diagnosis and remediation computational platform that will also be available on-line;
• Publications and talks in both national and international conferences.