An Analysis of the Development of Preschoolers’ Natural Science Concepts from the Perspective of Framework Theory

Abstract

The aim of the study was to explore children’s learning of natural science, and the authors chose a literature review as the means to achieve this aim. As the research into children’s emerging science is fragmented into many different theoretical perspectives and many kinds of studies are included, research results deriving from the field can be difficult to summarize. To resolve this issue, Vosniadou’s framework theory was used as the tool for data analysis. Results show emergence as a dynamic interaction between intuitive and counter-intuitive concepts leading to synthetic and scientific models in combination with developing epistemological and ontological skills. The accumulation of synthetic concepts seems to be the most common result of early formal educational input. Both intuitive and synthetic models are predecessors for scientific concepts and models, and these models as such are very important for emerging science. The conclusions suggest that diverse science experiences should bring about more synthetic models. This subsequently creates a wider basis for further development. Another important factor of emergence is the development of children’s epistemic skills and ontological shifts. Research results suggest that it is important for teachers to support children’s epistemic and ontological skills. Enhancing children’s understanding about causality and ontology is an important step towards developing formal concepts of science.

1. Introduction

Natural science and its introduction in early years education is a research field which has attained a lot of focus during the last four decades. One of the reasons for this increased attention is that preschool education has been shown to be important for socioeconomic equality [1], as preschools provide children with a diverse set of experiences, thereby supporting learning. Research on how children perceive and transform these experiences, i.e., children’s science or emergent science, has increased, and today we have a diverse array of knowledge concerning children’s understanding of natural science disciplines. The focus on early years education has so far mainly been directed towards biology and physics, as many of the phenomena within these topics can be experienced daily without tools to enhance our senses. Unfortunately, it is difficult to summarize the findings due to the theoretical diversity of the field, as different theories have provided a variety of perspectives for interpreting science education. This situation creates a need for analyzing results from different disciplines to paint an inclusive picture to understand better children’s science learning. The present article puts forth an analysis of children’s emergent science deriving from different theoretical fields by implementing and extending Vosniadou’s framework theory [2,3]. More specifically, this article presents an analysis of research literature concerning children’s concrete representations of natural sciences and explores “what” changes when children pass from one set of representations to another and “how” these changes occur.

1.1. Theoretical Background

The question of how children learn science and scientific concepts has been problematized by psychologists and pedagogues for centuries. The central questions based on which most of the learning theories can be evaluated are: what are the characteristics of children’s initial or everyday concepts, how do children acquire and use scientific concepts, and how do children generalize scientific concepts? The existing theories are inspired from a variety of philosophical standpoints, ranging from Piaget’s developmental theories, to Vygotskian, post-structural and cognitive approaches. The child’s learning progression is generally viewed as including both individual and social processes and being highly dependent on the pedagogical or educational setting [4].

Vygotsky’s notion of scientific concepts represented a step forward in educational psychology. He suggested that the social environment, and especially formal education, not only guides development, but is the crucial factor for the development of logical means of thinking [5]. This progression originates from making empirical categorizations of objects based on perceptual similarity and proceeds to identifying scientific concepts. The latter are characterized by the creation of a system of logically derived ideas as well as conscious awareness of thinking processes, and these are characteristics that permit children to explore and assess the content of experience and to hypothesize alternative explanatory systems. These developing functions provide the flexibility that is needed for grasping the methodological character of scientific thinking.

Piaget saw development as a series of stages: action (the child gains experiences by acting on the world), symbolic mediation (the child can make or make use of symbols), and then, later, abstract thought [6]. This description has since been expanded upon and now includes four different parts: (i) action, (ii) symbolic representations, (iii) functional dependency, and (iv) abstract thought [7]. Functional dependency is here used as a category to describe the stage when children practice making connections between reality and its representation [8]. In both cases, learning is seen as dependent on everyday experiences, and formal education is essential for this development [9]. Research today has become more subject-specific and now includes learning progressions of specific concepts, processes, and phenomena.

1.2. Previous Research on Children’s Learning Progressions

Research and analysis of children’s learning in science varies in the design. Some analyses derive from data that were collected at one point in time and can be seen as cameo moments of children’s understanding at different times in their learning progression. These types of analysis are valuable since they show moments in different individuals’ learning progressions. Other studies are longitudinal, and as such, have two or more points of data collection measuring children’s explanations or descriptions of a phenomenon usually before and after a teaching intervention. Analysis of data deriving from these studies provides the means to study the process of subject-specific learning, i.e., actual micro changes. The described learning pathways are usually either specific for one concept or phenomenon, such as light [10], or more general, suggesting learning pathways for a subject area [11]. The most basic learning pathways usually describe three different stages: everyday understanding, a combination of everyday understanding with some addition of formally introduced knowledge, and the third stage is usually a “scientific understanding”. Often, this pathway is seen as the result of the progressive introduction of scientific fragments. So, scientific emergence can be understood as a quantitative process, where children progressively reproduce more and more parts and aspects of the full scientific model.

One example of a subject-specific learning progression was provided by Malleus and Kikas [12], who performed a teaching intervention concerning the water cycle and found in their analysis the three categories: everyday knowledge (a combination of visual and perceptual knowledge), synthetic knowledge (a combination of everyday knowledge with scientific knowledge), and scientific knowledge. Their results also showed that the change caused by the intervention was mainly an accumulation of new synthetic concepts. Ravanis, Koliopoulos, and Boilevin [13] also collected data from a teaching intervention with the aim to explore preschoolers’ understanding of friction. The change observed from this intervention was described as children moving from everyday concepts to hybrid concepts to scientific concepts (no progress, partial progress, to progress). Similar words were chosen by Hannust and Kikas [14], who focused their study on astronomy and used the following terms to describe the children’s knowledge: no knowledge, descriptions (only describes the visible or personal experience), synthetic (combines abstract and visual concrete experiences), and scientific (accurate) in their categorizations of children’s answers. Other important conclusions drawn by this kind of research are that older children hold more synthetic and naturalistic conceptions than younger ones and that the importance of these two types of models for future learning cannot be underestimated [15].

This methodology has been expressed on a more theoretical level by Nobes, Frède and Georgaki, who, quoting di Sessa, state that “intuitive physics consists of a rather large number of fragments rather than one or even any small number of integrated structures one might call “theories” [16,17]. The authors argue that these pieces of knowledge are fragmented, lack coherence and are only activated under some specified circumstances. Τhis theoretical direction presumes that the passage from non-scientific to scientific thinking is an expression of the number of scientific elements that a child has learned and is using. As a result, the more scientific elements a child uses, the closer it resembles a modern scientific definition of a phenomenon.

Analyzing scientific change based on the number of scientific fragments that a child can reproduce per field, however, creates a situation of fragmentation. Science learning is much more than the aggregate of specific learning trajectories. To address this challenge, the present article offers an extension of Vosniadou’s theoretical system in preschool science didactics articles from different disciplinary and theoretical origins.

1.3. Vosniadou´s Framework Theory

1.3.1. Intuitive and Counter-Intuitive Concepts

Vosniadou’s perspective, named the “framework theory” [2], is based on studies of children’s scientific emergence, suggests a hierarchical structure that includes the notions of concepts, models, ontology, and epistemology (See

Table 1. Definitions of core concepts under the Framework Theory.

Concept	Defined as the specific representation that a child has regarding a process or an object. Concepts can be divided into two different kinds, intuitive concepts and counter-intuitive concepts. Intuitive concepts reflect children’s immediate experience and sensory input and are characterized by an initial epistemological and ontological level. Counter-intuitive concepts are generally derived from education and reflect more intrinsic characteristics, which cannot be derived from direct sensory input.
Model	Defined as a child’s full understanding of a process, the sum of the concepts and the epistemological and ontological principles which children use for explaining a phenomenon. Initial models are comprised of intuitive concepts and their frameworks. Synthetic models are characterized by a combination of intuitive and counter-intuitive concepts and early and mature epistemic and ontological skills. Scientific models are comprised of counter-intuitive concepts and mature epistemic and ontological skills
Ontology	Ontology refers to the set of more general ideas on the nature of reality. These ideas function as a set of expectations organizing the reality into groups of objects and processes. For example, physical ontology projects onto processes the qualities of everyday physical objects, whereas psychological ontology perceives processes as animate entities. The ability to flexibly deconstruct a previous ontological understanding and create a new one based on the information provided is evaluated as a higher ontological skill.
Epistemology	Epistemology refers to the causative mechanisms (the nature of knowing) used to describe and explain a phenomenon. The tendency to identify things with their external appearance is evaluated as an initial epistemic skill. The ability to create multiple representations, to challenge and evaluate one’s own skills and knowledge as well as others are understood as examples of mature epistemic skills.

and Figure 1). Within this framework, a concept is seen as the specific representation that a child has regarding a process or an object. Concepts can be divided into two different kinds, intuitive concepts and counter-intuitive concepts. Models are defined in a broader perspective as a child’s understanding of a process, the sum of the concepts and the epistemological and ontological principles which children use for explaining a phenomenon. Vosniadou’s theory makes some crucial adjustments to the theories that understand scientific emergence as a mere process of steady accumulation of scientific fragments. Changes in a specific concept of a process are important milestones, but learning is also a process of qualitative changes from one type of model to another. These models or frameworks are based on broader ontological and epistemological principles which guide how the child interacts with reality. Ontology refers to the set of more general ideas on the nature of reality, whereas epistemology refers to the causative mechanisms used to describe and explain a phenomenon [18]. Therefore, a change of an individual concept is interrelated with a broader structural change in a framework of ideas.

1.3.2. Initial Epistemological Level and Intuitive Concepts

Within this theory, intuitive models are seen to be comprised of intuitive concepts (see Figure 1). Intuitive concepts are characterized by an initial epistemological level (the nature of knowing) that is sensory-based. Here, interpretations and inferences are guided by immediate experience such as touch and visual sensory input. At this initial level, things are identified by their external appearance [19], such as, understanding what a cat is by using its characteristics, or seeing it as disappearing when it is no longer visible. Another example of the importance of visual input is that it is not uncommon for young children to make connections between different processes or objects just because they exist in the same visual context.

An example of how touch may affect categorizations can be found using the different states of water. For young children, the solid and the liquid states are often seen as different substances, since the solid state is often categorized as cold and the liquid state is connected to movement or as being wet [20]. This initial epistemology is the basis for further development, which is guided by formal education. Unfortunately, many of the concepts, classifications and causalities within the natural sciences are counter-intuitive, as they are derived from intrinsic characteristics, that is, characteristics that cannot be derived from direct sensory input. This initial epistemological level may become a hindrance for the development of more scientific concepts. Some of the effects on science learning that intuitive concepts, and the epistemology that they are based on, may have, have been described in literature, and these include a possibility of conceptual fragmentation [19], since the search for similarities between objects and processes does not initially move beyond the directly visual, a phenomenon that may act as a constraint for more abstract thinking.

1.3.3. Ontological Frameworks Connected to the Intuitive Concepts

Intuitive concepts are connected to initial ontological frameworks. These frameworks include a set of expectations of behavior and attributes. As new experiences are interpreted based on these frameworks, the expectations are transferred to the new experience. Once suggestions of initial ontological frameworks are defined as physical, psychological, mathematical, and lexical, these frameworks are applied to different ontological groups. “For example, physics applies to objects, psychology applies to animate entities, mathematics to numbers and their operations and language to lexical items and their operations” [21].

In the physical ontological framework, the expectations are the same as for everyday physical objects. For example, children expect that all objects should follow an up/down paradigm of gravity. When formally introduced ideas are interpreted as physical objects, then they include up/down gravity. This means that when objects such as an atom or a planet are placed within this ontological framework, then they are presumed to fall unless supported, just like everyday physical objects [18].

In the psychological ontological framework, children expect that objects follow the same rules as human beings, for example, having human intentions. If an object is placed in this group, then intentions are seen as the cause behind events. In this manner, placing an object in an ontological group may have consequences on what type of causation is connected to the object.

“...Each area is also governed by a distinct system of principles and rules of operation. Physical entities obey the laws of mechanical causality as opposed to psychological entities that are governed under intentional causality. Language and mathematics have their own unique rules and principles of organization” [21].

1.3.4. Synthetic Models

The emergence of scientific thinking depends on being able to use a scientific definition actively or to use fragments belonging to it. As previously mentioned, scientific ideas are formally introduced and often counter-intuitive. When children first learn a counter-intuitive concept, they form intermediate models where different types of explanations are combined. These synthetic models are created by combining intuitive and counter-intuitive concepts. The notions of misconceptions, alternative conceptions, or precursor models, which are also found in the research literature, are closely related but not identical to synthetic models [3]. Early synthetic models are indeed a part of a learning progression, and research results show that these synthetic models may simultaneously hold two conflicting explanations of the same phenomenon. Formally introduced explanations are used together with intuitive explanations. Explanations may also be used in a non-consistent way where the individuals have preferences towards specific types of problems. Another kind of synthetic model is created when children merge intuitive and counter-intuitive concepts to create their own new and original explanations. Unfortunately, if the intuitive conceptions are too far from the scientific material, learners might likely ignore the scientific explanation and instead favor the initial sensory-based existing ideas [22]. An example of a synthetic model of the Earth is when children portray it as a disk. This idea stems from children’s attempt to reconcile their intuitive physical ontology of up/down gravity, which stems from everyday objects, and the counter-intuitive concept that the Earth is a spherical ball. By combining these ideas, children often see it as a disk [3].

1.3.5. Scientific Models

A scientific model is defined as a full usage of all the scientific elements of a scientific definition and adequate epistemic and ontological capabilities. Being able to use a scientific explanation across similar, but different examples, and of the same phenomenon in a consistent and coherent manner, is a strong indication of scientific emergence.

1.4. Learning Progression as Moving between Models

Learning science is not just a process of replacing intuitive concepts with accurate scientific ones. It is a process where the initial ontological and epistemic frameworks become challenged, and new principles are incorporated into existing intuitive concepts (see Figure 2). Children’s initial ontological categories are replaced by more nuanced and complex ones.

Some examples of this change are the change from understanding the Earth as an everyday physical object, based on an up/down model of gravity, to viewing it as an astronomical object, which follows the specific laws of astronomy. This change represents an ontological shift as children form a new order of matter, which is characterized by its own specific types of laws and interpretations. This change also represents an epistemic change as non-visual factors, such as magnetic fields or rotation, are integrated in their definition of the Earth. Another attribute of an ontological shift is when children start to differentiate things which they thought were the same thing, for example, weight, volume, mass, and density [23]. This re-categorization under a new ontological order is a prerequisite for utilizing model-specific laws when trying to solve a natural science problem.

Other epistemic changes are making internal representations without external help [24], creating multiple representations of a phenomenon [25] based on its different relations, and evaluating and challenging their own misconceptions, beliefs, conflicting data and non-plausible conclusions [26]. Also, to test hypotheses and “to differentiate between ideas, experiments, and results” [15] are epistemic changes. Scientific emergence depends on this dynamic process for epistemic and ontological development, as synthetic models are important intermediate bridges towards the scientific model. Through this process, children begin to master scientific fragments and to recognize the discrepancies of their initial ontological and epistemological beliefs. They have an opportunity for changing their general conceptions, which eventually facilitates scientific thinking. Using a scientific explanation across similar, but different examples, and cases of the same phenomenon in a consistent and coherent manner, is a strong indication of scientific emergence.

2. Materials and Methods

Method

To achieve the goal, a literature study was performed on the 10th of September 2021 using ERIC with the words “preschool natural science” as the topic. The search resulted in 200 articles. From these, the papers which focused on a different educational level or population, like the parents and teachers, were excluded. Since the focus of the analysis was children’s conceptions of natural science processes, articles which focused on experimental pedagogic practices were mostly excluded. The result was 53 articles remaining, and out of these, 28 articles were chosen focusing on theoretical variation and diverse natural science content. The fields which were included were: children’s perceptions of the Water cycle and Cloud formation [27,28], Evaporation [29], Light [10], Heat [30], Shadow [31], Lunar phases [32], Thermometer [33], Contraction and expansion (iron heated ball) [34], Combustion (candle) [35], Friction [13,36], Animal taxonomy/animal-environment specificity/adaptation [11,37,38], Aliveness [39], Fungi [40], Germs [41], Environmental biology [42], Gardening and early math [43], Understanding of food [44], Conceptions of Water, Molecules, and Chemistry [45], Molecules and atoms [46], Rocks [47], Water Physics [48], Sound in space [49] and Density [23]. The first article included was published in 1985, with an increasing number of published works after 2010 (See Figure 3).

The empirical results of the articles were categorized under intuitive, synthetic and scientific models (see

Table A1. The full analysis of Intuitive and Synthetic models using Vosniadou’s framework theory.

Content Article nr	Intuitive Models (Intuitive Concepts)	Classification	Synthetic Models (Intuitive and Non-Intuitive Concepts)	Classification
Water cycle and Cloud formation [27,28]	Water cycle is water that moves around.	Ees	Static model of rain (clouds bring water but rain comes from the sky, clouds may be made of some substance other than water like cotton or something soft)	Sf: reflecting aspects from the scientific definition of water cycle Ees–cloud like cotton
	Condensation is droplets on the window.	Ees	Usage of the concept’s atmosphere, temperature	Sf: using scientific vocabulary
	Rain and clouds are seen as separate objects	Ees	Rain is created from the collision or shaking of clouds	PoSf: mechanical causality connected to the water cycle
	Rain disappears after falling.	Ees	Clouds are water in steam/vapor form made by droplets and create rain	Sf: using scientific vocabulary Mes: interconnecting clouds, vapors and process of creation of rain
	Clouds are weather events	Ie: Everyday associative knowledge
	Clouds are seen as solid objects	Po
	Understanding rain as a super-natural agent	Pho -attribution of purposive intentionality to the clouds
Evaporation [29]	Water disappears	Ees: -water disappears when immediate visibility is lost	Water moves to an alternative location “in the sky, sun, ceiling, air, or clouds.”	Sf: recognizing counter-intuitive tendency of water moving upwards CnOc: The idea that vapor moves upwards presupposes a new ontological category of a type of matter which goes upwards defying the behavior of everyday physical objects due to gravity
	Water is absorbed in the floor or the ground.	Ees	Water vapors are small drops that are scattered in the air.	Sf: recognizing counter-intuitive tendency of water moving upwards. CnOc: The idea that vapor moves upwards presupposes a new ontological category of a type of matter which goes upwards defying the behavior of everyday physical objects due to gravity
			Water is formed by oxygen and hydrogen.	Sf: using scientific vocabulary
			Water vapor is a mixture of ‘water with heat’ or of air and water. Fire and air were viewed as “movers”, which prompted this “unnatural behavior of water”	Mes: connecting different parts of the experiment for understanding the transformation of water into vapor PoSf: understanding of the basic elements as mechanical powers
Light [10]	Cannot be in space, can only be found in lamps or illuminated spots	Ees: light disappears when immediate visibility is lost	Not only identifying light in lighted area but also in space	Mes: understanding movement of light as related with light source, space, and lighted area
	Correlate light with heat	Ees Io: identification but no differentiation
	Light is an autonomous entity.	Pho: purposive intentionality
Heat [30]	Heat is the feeling of burning	Ees: heat is its immediate visible/tactile characteristic	Heating and cooling are the result of difference in temperature due to energy transfer	Sf: Recognizing aspects from the scientific definition Mes: Understanding heat change as an expression of energy transfer
Shadow [31]	Shadows are autonomous substances or entities.	Pho: -shadows have intentions Ees: identification of shadows with their immediate visible/sensual characteristic	Shadows are produced when an object blocks a source of light, and the light follows a straight path	Mes: Understanding shadows as the result of the interrelation between light source and blocking object
Lunar phases [32]	The moon’s light is a light which shines through the moon. Its source is not the sun.	Ees: -identification of lunar light with its immediate sensory characteristics
	Shadowed and the lighted areas of the moon are two different things.	Ees
	Animistic understanding of the moon	Pho: purposive intentionality connected to the moon
Thermometer [33]	Generally associating the thermometer with fever and the feeling of being ill	Ie: Everyday associative knowledge	Thermometer measures the change in temperature, numerical increase is a sign of heat fluctuation. The latter is a sign of fever, illness.	Sf Mes: connecting the thermometer’s function with temperature, its numerical increase, fever, illness and feeling of internal heat
			Thermometer’s changes were understood as some kind of mechanical force pressing the mercury.	PoSf: understanding of mercury’s movement as mechanical power
Contraction and expansion (iron heated ball) [34]	The iron expands due to its own essence and not because of heat	Ees: expansion is attributed to an unexplainable inner cause	Identifying contraction and expansion with heating and cooling/temperature fluctuations determine expansion and contraction of iron.	Mes: connecting contraction and expansion with heating and cooling
Combustion (candle) [35]	The candle burned out because it wanted to, and because it was predetermined to do so.	Ees: connecting the burning out with visually accessible traits. Pho	Using the words condensation, lack of oxygen/air and moisture	Sf Mes: connecting combustion with condensation, lack of oxygen/air or vapor
	Attributing the burning out due to the temperature within the vessel, the air/wind, the melting of the candle, change of color, emission of light or appearance of stain	Ees: connecting the burning out with visually accessible traits.
Friction [13,36]	Children cannot explain why some moving objects roll closer or farther from each other	Pho: roughness and steepness are not differentiated from the rest of visual characteristics of the objects	Understanding friction through roughness and steepness of the surfaces (smoother or rougher) and the characteristics of the weights of the moving objects (heavier or lighter)	Mes: connecting friction effects as an expression of the surfaces and the weights of the rolling objects
			Understanding friction through roughness and steepness depends on being able to observe, describe, experiment predict, compare, contrast,	Mes: Synthetic models of friction are mediated by mature epistemic skills
Animal taxonomy/animal-environment specificity/adaptation [11,37,38]	Taxonomy based on concrete, observable animal traits/Grouping species by physical appearance.	Ees	Connecting specific anatomical characteristics responding to environmental changes and evolutionary challenges	Sf Mes: classifying animals based on more complicated criteria, like environmental traits and characteristics
	Members share an underlying and immutable nature (essentialism)/Animals do not vary or change/no understanding that evolutionary change happens over many generations.	Ees: All members are represented as one singular and unchanging pictorial representation	Preliminary understanding of adaptation that animals change to develop beneficial traits for themselves or as respond to a need, defined as transformationalism	Mes: interconnecting traits with environmental challenges PhoSF
	Organism-environment specificity without knowing why.	Ees: connecting organisms to their environment by visual co-occurrence
	Children apply human characteristics to animals/anthropomorphism.	Pho
	Knowledge of animal-forest specificity was a collection of isolated concrete facts.	Ees: children’s ideas as recollections of visual characteristics
Aliveness [39]	Moving objects are alive. (a toy-car, a river is alive. A plant is not alive)	Ees: identification of aliveness with immediate visibility, movement	Aliveness is evaluated based on having biological functions, dependence on the environment (the toy-car and the river are not alive, whereas the plant is alive)	Sf: recognizing biological functions for evaluating aliveness Mes: understanding aliveness as interrelated with biological functions
			The human body as system of internal organs with different functions	Mes: understanding aliveness as interrelated with biological functions
Fungi [40]	Life was only attributed to plants and animals and not fungi	Io: identification and not differentiation of fungi from other animals and plants.	Awareness of fungi and differentiation from other animals and plants	CnOc: creation of a new ontological category of fungi
			Recognizing fungi was correlated with cognitive and linguistic development such as being able to observe, pay attention and compare data as well as to understand and communicate complicated verbal instructions	Mes: Synthetic models of categorizing nature are mediated by mature epistemic skills
Germs [41]	Related with everyday problems regarding health and sickness/that they are not alive and only bad.	Ees	Recognizing biological functions/ they are alive, that they perform good functions.	Mes: understanding germs as interrelated with a biological function
	Anthropomorphism	Pho: connecting purposive intentionality to germs
Environmental biology [42]	Children cannot identify fractal patterns of trees	Ees: no analytical skills, does not abstract common traits	Being able to recognize fractal patterns of trees.	Sf: recognizing aspects from the biological structure of trees
	Not being able to recognize the fractal structure of trees was related with natural/spontaneous observation skills, difficulties noticing, observing patterns, and expressing observations	Ees	Recognizing fractal patterns of trees as related with being able to collect, process, measure and classify their data, to use inquiry, to have argumentation skills and interpret the data and to share and communicate their findings with their peers	Mes: Synthetic models of environmental biology are mediated by mature epistemic skills
Gardening and early math [43]	Not being able to prepare, plant and harvest the garden beds	Ees	Being able to prepare, plant and harvest the garden beds,	Sf: learning core practices of gardening
			Having conscious awareness of gardening practices and ecological principles, using correct words to specify mathematical concepts in relation to gardening.	Mes: connecting traits with environmental challenges
			Being able to observe, predict, evaluate, compare and use number-related concepts (addition and subtraction, fractions), spatial orientation, and size estimation and comparison)	Mes: Synthetic models of gardening are mediated by mature epistemic skills
Understanding of food [44]			Finding, weighing, measuring, sorting various traits of foods, identifying, and creating a meal based on healthy criteria/examining food books and pictures.	Sf: core practices of food processing, scientific vocabulary
			The above skills were related with problem solving, exploration, creativity, literacy, and inquiry skills	Sf Mes: Synthetic models of understanding food are mediated by mature epistemic skills
Conceptions of Water, Molecule, and Chemistry [45]	Water is how it appears/concretely, as something to drink/difficulty providing a verbal explanation of what water is.	Ees	Emerging understanding and usage of scientific words, like phase change, surface tension and the water cycle	Sf
	Understanding of molecules as concrete everyday physical objects/metaphoric representations of molecules are understood as literal.	Po: molecules are understood as everyday physical objects	Water is part of a broader system, specifically the physical world.	CnOc: Water is understood as part of a broader natural system as a new ontological category
	The water molecule has properties such as being blue or soft.	Ees	Water is part of a broader system, specifically the physical world.	CnOc: water is perceived as being part of a broader ontological category
			Being able to understand the relation between temperature and the speed of molecules.	Mes: molecular behavior as interrelated with temperature and speed
Molecules and atoms [46]	Interpreting atoms as meatballs	Po	They just look like meatballs, they are everywhere in everything	Sf
Rocks [47]	Rocks are understood through their observable properties such as color, size, shape, and weight, hardness, and luster	Ees	Understanding rocks through the concepts of erosion, strength, hardness, and weathering	Sf
			The above skills were related with descriptive and drawing abilities, summarizing their learning, and argumentation skills	Sf Mes: Synthetic models of understanding rocks are mediated by mature epistemic skills
Water Physics [48]	The water will always flow because there exists a pipe.	Ees	Recognizing that the water’s pressure is due to the height difference between the reservoir and the pipe’s exit, the hole size and the resistance along the path of flow	Sf: Aspects from the scientific definition of water physics
	Inability to predict flow of the water.	Ees Ie: children could not apply cause and effect because they could not see	The above skills were related with the ability to create, evaluate, distinguish, and connect artifacts, identifying goal, exploring, planning, analyzing a problem, coordinating the effect of multiple factors, comparing evidence, and finding common features.	Sf Mes: Synthetic models of understanding water physics are mediated by mature epistemic skills
			Understanding water physics based on gravitational or energy restrictions	PoSf: water systems as mechanical systems
Sound in space [49]	Sound is identified with the objects which produce or receive it/connect sound with its source or its context.	Ees	Recognizing sound independently of sources of production, its origin, or forms such as musical sound, recognizing the presence of sound between a source and receiver, understanding that sound is a non-material substance, sound as invisible or transparent as a ghost or smoke, sound is understood through the notions of ‘vibrating’ and ‘resonating’.	Sf
	Sound is exclusively related with voice.	Ees	Sound as an autonomous material entity, sound is not identified with a medium for its propagation, it can propagate through empty space between the particles of a medium irrespectively of whether it affects them or not.	CnOc: understanding the microscopic definition of sound
	Attributing material properties to sound, a ‘solid’ object cannot go through another one, it is impossible for sound to move across solids	Po: -sound is understood as a physical object	The wave model refers to sound as a vibration within a material medium essential for its propagation, the medium is made up of particles that oscillate around an equilibrium position, t is this oscillation that represents sound.	CnOc: understanding of a microscopic definition of sound. Mes: sound production related with the oscillation of particles
Density [23].	Objects sink because they are made of something heavy, objects float because they are light, density is not yet differentiated from weight, volume, and size.	Ees Po: -Identifying high density with large weight is related with an observation that heavier objects have more gravitational powers.	Some objects float due to their relative weight in comparison with the surrounding water	Mes: understanding relative weight by interrelating the weight of the different objects
	Some elements float whereas others sink due to the inner characteristics of the substance	Ees	Differentiating among volume, weight, size and density	Sf: Recognizing aspects from the scientific definition of density

in Appendix A). The data categorized under the intuitive category included both intuitive models and concepts. Identifiers for this category were expressions of early epistemic skills, a physical object, or a psychological ontology without the identification of a scientific fragment. Also, data that followed a sense-based understanding and practice towards a process were also identified as intuitive. For example, studies inspired by the Framework theory and focusing on Mathematics education categorized preschooler’s tendency to understand numbers discretely as intuitive thinking. At this stage, even logical numbers such as fractions or decimals were seen as discrete numbers [50].

On the other hand, all data showing counter-intuitive concepts and synthetic models were designated as belonging to the synthetic category. Identifiers for this category were fragments from scientific definitions, the presence of scientific vocabulary, and mature epistemic skills. Representations which exhibited a differentiated or an intuitive (physical or psychological) ontology with the presence of scientific fragments were also categorized as synthetic. Similarly, in Mathematics education, the presence and usage of logical numbers signifies the emergence of synthetic mathematic concepts. Finally, the science category included the data that contained a full scientific definition of a process.

3. Results

Analysis of the articles was carried out using the categories of intuitive, synthetic, and scientific to classify the empirical data from the theoretically diverse field. The abbreviations used in the analysis are defined in

Table 2. Explanation of abbreviations used in the analysis.

Abbreviation	Definition
Ees	early epistemic skills, identification of a process or object by its visible/sensory traits
Ie	Intuitive epistemology
Io	Intuitive ontology
Sf	Scientific fragments
Pho	Psychological ontology—Intentions without scientific fragment
PhoSf	Psychological ontology—Intentions with scientific fragment
Po	Physical objects ontology—without scientific fragment
PoSf	Physical objects ontology—with scientific fragment
CnOc	Creation of new ontological category
Mes	Mature epistemic skills
Sd	Scientific definition

.

The definitions were then used for classifications within the framework theory, into the categories of intuitive, synthetic and scientific; both concepts and models were merged into these categories. The full analysis (Table A1) is presented in Appendix A. Due to the size of Table A1, only a part of the table is presented in the Section 3 (

Table 3. Categorizations using Vosniadou’s framework theory.

Content Article nr	Intuitive Models (Intuitive Concepts)	Classification	Synthetic Models (Intuitive and Non-Intuitive Concepts)	Classification
Water cycle and cloud formation [27,28]	Water cycle is water that moves around.	Ees	Static model of rain (clouds bring water but rain comes from the sky, clouds may be made of some substance other than water like cotton or something soft)	Sf: reflecting aspects from the scientific definition of water cycle Ees–cloud like cotton
	Condensation is droplets on the window.	Ees	Usage of the concept’s atmosphere, temperature	Sf: using scientific vocabulary
	Rain and clouds are seen as separate objects	Ees	Rain is created from the collision or shaking of clouds	PoSf: mechanical causality connected to the water cycle
	Rain disappears after falling.	Ees	Clouds are water in steam/vapor form made by droplets and create rain	Sf: using scientific vocabulary Mes: interconnecting clouds, vapors and process of creation of rain
	Clouds are weather events	Ie: Everyday associative knowledge
	Clouds are seen as solid objects	Po
	Understanding rain as a super-natural agent	Pho -attribution of purposive intentionality to the clouds
Evaporation [29]	Water disappears	Ees: -water disappears when immediate visibility is lost	Water moves to an alternative location “in the sky, sun, ceiling, air, or clouds.”	Sf: recognizing counter-intuitive tendency of water moving upwards CnOc: The idea that vapor moves upwards presupposes a new ontological category of a type of matter which goes upwards defying the behavior of everyday physical objects due to gravity
	Water is absorbed in the floor or the ground.	Ees	Water vapors are small drops that are scattered in the air.	Sf: recognizing counter-intuitive tendency of water moving upwards. CnOc: The idea that vapor moves upwards presupposes a new ontological category of a type of matter which goes upwards defying the behavior of everyday physical objects due to gravity
			Water is formed by oxygen and hydrogen.	Sf: using scientific vocabulary
			Water vapor is a mixture of ‘water with heat’ or of air and water. Fire and air were viewed as “movers”, which prompted this “unnatural behavior of water”	Mes: connecting different parts of the experiment for understanding the transformation of water into vapor PoSf: understanding of the basic elements as mechanical powers

). The statements in columns 2 and 4 in Table 3 and Table A1 are the original findings from the referenced articles as cited in column 1. These findings were re-categorized under the Framework theory as intuitive, synthetic or scientific concepts and models. The analysis was based on the empirical data of a variety of studies with different theoretical perspectives. The framework theory afforded a new context for re-interpreting the diverse perspectives.

3.1. Intuitive Conceptions and Models

Results indicate that children’s initial ideas were based on their immediate senses. For example, the water cycle was understood as moving water, evaporation as the disappearance of water, light being identified with illuminated areas and shadows with shadowed areas, the light and shadow of lunar phases seen as different things, lack of combustion seen as caused by coldness or wind, animal taxonomy based on observable features and aliveness identified with motion. When children were asked to provide an answer to these problems, the answers included everyday experiences. Similarly, a recurrent finding was that objects stopped existing when children’s visible connections with them were obstructed [12]. In all these examples, the concrete content of immediate senses and impressions was shown to be the basis for children’s description of the process.

The results also show that objects were identified with their everyday uses and function or were associated with other elements in the immediate vicinity of the object itself. For example, the thermometer was generally associated with fever, germs with sickness, aliveness with movement, etc. Often the children could not explain how they made these spontaneous associations. Accordingly, the results show that children’s initial conceptions of scientific processes were seen as caused by specific, sensory-based, or visibly accessible, macroscopic objects along with their uses and properties.

The analysis identified two types of ontological structures: the psychological and the physical object context. The psychological ontology places the cause of a certain phenomenon as purposive intentionality, meaning that when children were using the psychological framework, they provided explanations based on the object’s own intentions. For example, clouds, light, shadows, the moon, the candle, animals, and germs were all attributed with having purpose. In the second ontological framework, called the physical object ontological framework, children applied the properties of everyday objects to scientific concepts. When children tried to solve problems regarding clouds, water, molecules, sound, and density, they treated the physical objects as being alive. For example, children tended to identify the sinking of highly dense objects with large weight. This was related with the everyday observation that heavy objects are difficult to lift. In this sense, the everyday gravitational model of physical objects was used for solving problems regarding density. These results suggest that, initially, new scientific experiences are integrated into preexisting everyday concepts instead of being differentiated into new concepts.

3.2. Synthetic Models

The synthetic models derived from the analysis included some scientific fragments, notably fragments of a formal scientific explanation of a phenomenon and usage of scientific vocabulary. Although the usage of a specified scientific vocabulary or a specific practice might not always correspond to a full understanding of its meaning, it suggests the emergence of a more complex understanding of a phenomenon.

These conceptions were characterized by interrelating sensory input in new ways. When children included counter-intuitive concepts as points of reference, they were able to make more complex associations between different and visually distinct situations. For example, children began interconnecting rain, clouds and evaporation (rain cycle), water, boiling and vapor (evaporation), heating with energy (heat), shadows with light sources and blocking objects (shadows), the thermometer with the mercury’s ability to measure temperature through its numerical increase, fever, illness and feeling of internal heat (thermometer). Contrary to their everyday representations, characteristics belonging to different visual frameworks were correlated, providing a path to new types of causal explanations. Children started to differentiate scientific problems from the elements of their immediate surrounding environment and began to view them with more complex logical interrelations. This represents a more mature epistemic skill as children become emancipated from the phenomenalism, the tendency of identifying something with how it appears. Viewing things as interrelated permitted the children to understand non-visual properties of elements, such as density. These non-visual, logically derived, properties then progressively become the determining characteristic of concepts. One example is the development of children’s understanding of animals, as they shift from using their external characteristics as identifiers towards using their biological functions as identifying criteria.

These new synthetic models and counter-intuitive concepts (from intuitive to counter-intuitive concepts) and epistemic changes are interconnected with some ontological shifts. Previously undifferentiated attributes and characteristics become insufficient for providing full explanations, and more aspects of the scientific definition are learned. For example, children start to understand that density is different from weight, sound from the tool that created it, or the microscopic level from the macroscopic. These are new ontologies, and they are characterized by a new set of attributes and causative explanations.

In many cases, the introduction of new scientific fragments is not accompanied by a broader ontological shift. Sometimes children interpreted the new fragments by using incompatible ontological models, mainly as physical objects, or as psychological entities. For example, in the children’s understanding of rain formation, the children used the collision model, which states that rain is created from the collision between or shaking of clouds. Also, the thermometer’s changes were attributed to a mechanical force pressing the mercury [33]. The behavior of the mercury is viewed as an object which is physically pushed by some other object. The analysis also identified a psychological framework, interpreting animal adaptation as a subjective change.

In these models, natural processes are viewed as mechanical or purposeful, indicating a physical or a psychological ontological framework. Mechanical thinking was related with the noted macroscopic character of the children’s everyday concepts. In the same way, providing an explanation to a scientific problem by attributing purpose to it, means treating objects as human beings. It seems that these initial models are compatible with counter-intuitive concepts. These are adequate for including certain scientific fragments, before creating a new and scientifically adequate ontological category. Learning more counter-intuitive concepts causes ontological frameworks to differentiate into new ontological structures.

A few scientific models were noted in four articles (See

Table 4. A summary of the scientific models found in the analysis.

Content Article nr	Scientific Models	Classifications
Water cycle and cloud formation [27,28]	Understanding all the necessary scientific fragments of the water cycle. Specifically, that rain is water, clouds consist of tiny water droplets or ice crystals, from clouds, rain water does not disappear when it hits the ground, and that rain water evaporates and becomes a cloud when condensed.	Sf
Molecules and Atoms [46]	Understanding that all things are made of smaller particles, and the phenomena of evaporation, filtering, dissolving, stirring, and mortaring as well as being able to use the magnifying glass, the chromatographer and have some basic understanding of the microscope.	Sf
Sound in space [49]	Defining sound as an oscillation of molecules in an elastic medium (such as air). This oscillation mechanically stimulates its neighboring molecules, in a process leading to the creation of a sound wave which eventually leads to the stimulation of the receiver’s ear.	Sf
Density [23]	In the most formal sense, an understanding of buoyancy would involve knowledge of relative densities of substances and liquid media, weight, volume, surface area, surface tension, and so on.	Sf

). These models were not necessarily identified in children’s representations but were formally defined by the researchers as ideal full models.

4. Discussion

The previous studies presented here are each significant contributions that characterize children’s scientific emergence. They offer concrete examples of how children’s scientific thinking develops in various fields, which can be used by practitioners for structuring educational interventions. The present study attempted to maximize the contribution of these articles, by not viewing them as individual contributions, but rather in their mutuality and under a common explanatory system.

To accomplish the purpose of this study, it was necessary to separate the specific findings of the articles from their theoretical explanations and attempt to see whether Framework theory could incorporate them in its own explanatory system. For example, in the research regarding combustion, researchers adopted a theoretical framework which used scientific, synthetic, naturalistic (agentive and non-agentive) and non-naturalistic explanations. These categorization schemes did not make use of ontological and epistemic criteria; for example, a scientific concept was defined as “compatible with the current state of scientific knowledge” [35]. Synthetic categorizations were defined as being made of both scientific and non-scientific alternative elements, as a synthesis of scientific and non-scientific fragments. These synthetic and scientific categorizations were re-interpreted as synthetic concepts and models, due to the existence of scientific fragments and epistemic skills. Scientific categorizations in the above-mentioned article were transformed into synthetic due to lack of data regarding epistemic skills expressed through these concepts. Non-naturalistic and naturalistic agentic explanations were re-interpreted as intuitive concepts having a psychological ontology, whereas all other naturalistic explanations were re-interpreted as intuitive concepts having early epistemic skills. Both were interpreted as intuitive because they reflected an unmediated sense-based relationship to the burning of the candle, mainly documenting what was visual.

A similar logic was adopted in the research regarding aliveness. According to the results in these articles, children’s initial ideas identify aliveness with movement, whereas more mature ideas include other criteria like biological functions and dependence on the environment. Here, scientific fragments are used to categorize children’s thinking. Following Framework theory, the ideas which identify aliveness with movement were re-interpreted as intuitive, given that children use only sense-based data. The more mature ideas necessitated a more challenging processing, interconnecting immediate visual data with the concepts of biological functions and environmental dependence. This capability reflects more mature epistemic skills, and in this sense these ideas were re-interpreted as synthetic.

By elaborating the conceptual categorization criteria of Framework theory, the present research was able to detect similar patterns across different natural science research and thus extract more results from them. The analysis was able to identify epistemic and ontological frameworks in children’s intuitive and counter-intuitive concepts. Children’s specific concrete ideas were based on a certain set of processing rules, through which they comprehended their experiences. Early epistemic skills were identified in almost all scientific disciplines examined, except in the case of the thermometer, friction, fungi, molecules and atoms. Children tended to investigate the explanations of a phenomenon in its immediate visual environment. For example, in the intuitive notion of aliveness, the latter is identified with movement. These immediate connections are contrasted with the types of associations that were identified in counter-intuitive concepts. In other words, children needed to disregard their senses-based interpretations and instead analyze phenomena systematically. This challenge was identified across almost all learning progressions.

The results show that it is important to explore children’s ideas about causality and to promote thinking based on mature epistemic skills. Understanding the interrelated character of scientific concepts rather than memorizing the scientific material is a process closer to scientific emergence. This finding suggests that when children learn scientific material, they make use of their epistemic and ontological skills. In this sense, scientific emergence is not only about learning the proper scientific idea in a specific discipline, but rather to learn how to treat their data in a scientific manner, how to think based on a scientific methodology. Consequently, teachers should not only teach children about scientific problems, but also the principles of scientific methodology and thinking. A curriculum which could introduce children to the concept of causality and the nature of knowledge could help children’s reasoning skills overall.

The analysis also identified the usage of an initial ontological framework with the presence of scientific fragments. This implies that psychological and mechanical models are important for understanding new ideas. This finding indicates that these different fields do not exist in their individuality and separateness, but some are used as prototypes for new learning. Thinking through analogies could become a valuable tool for supporting children’s understanding of science and encourage their development of epistemic and ontological skills as they process their experiences from different contexts [51].

Based on these results, Vosniadou’s framework identifies more ways of analyzing data from natural science didactics. More importantly, it creates a methodology which could move beyond the theoretical multilingualism and especially the fragmentation of the field of natural science didactics. Framework theory moves beyond a domain-specific logic of emerging science, which presumes that the developmentary course is individual in each specific field [16]. Moreover, it leads to different practical educational consequences, in contrast with Nobes, Frède’s and Georgaki’s approach. Whereas the latter presume that children do not have any cognitive frameworks to hold them back from conquering a scientific model, Vosniadou states the importance of synthetic models. For Nobes, Frède’s and Georgaki, scientific emergence depends on whether the child’s environment will clearly describe scientific knowledge without misconceptions. On the other hand, for Vosniadou, working with the synthetic models, even if they are not scientifically accurate, will aid children’s concrete specific and general cognitive development.

Vosniadou’s perspective represents a step towards more Vygotskian approaches, in which scientific thinking is the result of a qualitative reorganization of psychological functioning. Vygotsky stated that spontaneous everyday concepts develop differently than scientific concepts. The first arise because of children’s direct interaction with actual objects and a psychological elaboration based on external similarity. On the other hand, scientific concepts express a mediated connection to the object through the creation of a system of logical relationships between concepts [52,53]. This process was observed in the present analysis when children started to use inquiry to analyze their natural science experiences.