Bridging Laboratory Inquiry and History of Science: Enhancing Scientific Literacy Through Explicit and Reflective Approaches to the Nature of Science

Abstract

This study proposes an innovative instructional approach to promote scientific literacy by integrating the Nature of Science and the Nature of Scientific Inquiry with experimental practice and the history of physics. The aim is to foster a deep understanding of how scientific knowledge is constructed and to promote informed trust in science. Using an explicit and reflective methodology, the intervention combines experimental activities with historical reflection. The core of the learning sequence is the experimental reconstruction of Galileo’s studies on falling bodies, based on the historical manuscript folio 116v, an original document that provides the empirical evidence for the law of falling bodies, illustrating the transition from raw experimental data to mathematical formalization. Through this activity, students engage with key epistemic aspects of scientific practice, including the management of uncertainty—distinguished into statistical/aleatory and structural/epistemic forms—the probabilistic nature of scientific knowledge, the predictive power of models and theories, and the underdetermination of scientific theories. Additional themes addressed include the role of thought experiments, the importance of communicating results for scrutiny and validation, the function of models as mediators between theory and phenomena, and the process of de-idealization. The study also challenges the persistent myth of a single, linear “scientific method,” highlighting instead the theory-laden character of scientific inquiry and the central role of the scientific community. This dimension is explored through the historical comparison between Galileo and Mersenne, which illustrates elements of the scientific ethos and the role of peer review as a mechanism for the correction and refinement of knowledge. The results obtained with pre-service teachers, with whom this instructional sequence was implemented, indicate that this contextualized approach facilitates the overcoming of a view of science as a set of absolute truths. Instead, it promotes a more mature understanding of science as a dynamic, provisional, and self-correcting human enterprise, while equipping future citizens with the critical tools necessary to navigate the challenges of the twenty-first century.

1. Introduction

In recent decades, a growing consensus has emerged among science education researchers that the transmission of scientific content alone is insufficient to achieve the broader aims of science education. Instead, there is an increasing emphasis on cultivating a deeper epistemological understanding of how scientific knowledge is constructed, how it interacts with societal contexts, and how to foster a well-grounded trust in science (TIS)—one that acknowledges science’s fallibility rather than presuming its infallibility.

This epistemic orientation is condensed in the frameworks of the nature of science (NOS) and the nature of scientific inquiry (NOSI), which collectively address the empirical, inferential, and provisional dimensions of science, as well as the self-corrective mechanisms of the scientific community. These frameworks stress not only the empirical basis and creative processes of inquiry, but also the importance of uncertainty and critical reflection in the construction and validation of knowledge.

The present study builds upon and extends traditional NOS frameworks by incorporating underexplored dimensions such as forecasting, model-based reasoning, and the interplay between theory and data. Forecasting is presented as a form of inference grounded in the recognition of patterns and trends, highlighting the predictive power of scientific models and theories. This predictive capacity is inherently linked to the provisional and self-corrective NOS, expressed through the continual revision and falsification of models.

Particular emphasis is placed on the correspondence principle and the concept of validity domains, which underscore how multiple theories can coexist by operating within specific empirical boundaries. Through this lens, students develop an understanding of model construction as both an empirical and theoretical practice, involving idealization and subsequent de-idealization processes. These activities are designed to support the development of practical and epistemological competencies related to the dynamic and cumulative nature of scientific knowledge.

The concept of uncertainty, often perceived as a limitation, is reframed as a defining and constructive element of science. Two forms of uncertainty are discussed: statistical uncertainty, which pertains to measurement precision, and structural uncertainty, which concerns the underlying relationships among variables.

To provide a rich and contextualized learning experience, the study incorporates historical case analysis—particularly the work of Galileo on falling bodies—as a means to explore theory-ladenness, the myth of a unified scientific method, and the collective nature of scientific practice. The contrast between Galileo and Mersenne serves as a point of departure to examine the institutionalization of peer review and the role of the scientific community in validating knowledge. While a comprehensive historical and epistemological reconstruction—incorporating the foundational contributions of figures such as Francis Bacon and René Descartes—is essential for a full philosophical treatment of the subject, such an exhaustive analysis or an overview of the development of scientific research lies beyond the scope of the present work. This study does not aim to provide a complete history of scientific thought; rather, it intentionally focuses on specific historical episodes that serve as pedagogical catalysts for eliciting student reflections on the NOS within a laboratory context.

The teaching intervention was developed and tested over three years with student and pre-service physics teachers. It combines the goals of laboratory-based learning with explicit reflection on NOS and NOSI elements. The instructional sequence is inspired by Galilei’s historical notes (Schvartzer et al., 2020), particularly folio 116v, and engages students in experimental design, modeling, and critical evaluation.

Through a carefully scaffolded inquiry-based learning process, students conduct experiments, analyze data, and construct models, which are then refined through cycles of prediction and empirical testing. This promotes an authentic experience of scientific practice, wherein the mediation of theory by models is made explicit, and the provisional status of knowledge is not only acknowledged but critically examined.

The culminating phase involves structured historical reflection, inviting students to interrogate scientific concepts in light of philosophical and historical debates. This reflective approach helps cultivate a more nuanced, critical, and trust-based view of science as a dynamic, uncertain, and profoundly human enterprise.

In summary, integrating experimental activities, historical enquiry and explicit reflection on the NOS is a significant innovation in promoting scientific literacy (Çilekrenkli & Kaya, 2022; Xiang & Han, 2023). Specifically, the Explicit–Reflective Teaching approach is widely recognized as a leading method of science instruction (Bugingo et al., 2024; Deng et al., 2025; Khishfe, 2023; Mulvey et al., 2016). Although the literature often lacks detailed information on specific classroom implementations, several well-established activities serve as foundational benchmarks. These include ‘black box’ tasks, which engage students in formulating hypotheses and evaluating evidence (Khishfe, 2023), and visual perception exercises, such as Tricky Tracks, which illustrate the socio-cultural and theory-laden nature of scientific knowledge. Furthermore, techniques such as interactive historical vignettes and the use of historical and autobiographical material have been shown to be effective in humanizing the scientific enterprise through structured in-class discussions. In many proposals, reflection is typically scaffolded through structured questions followed by whole-class discussions to support students in articulating and revising their conceptions of NOS. Furthermore, there has been increasing attention devoted to the use of narratives—both historical and autobiographical—as powerful tools to contextualize scientific practices and make NOS aspects more accessible and meaningful (Witucki et al., 2023; Tolvanen et al., 2013). Moreover, practice-based instruction in Socioscientific Issues has recently been shown to improve students’ NOS comprehension, particularly with regard to its social aspects (Leung, 2020).

Building upon these foundations, this work’s primary novelty lies in the development of a comprehensive framework that connects laboratory research with the history of science. Unlike typical approaches that treat NOS as an isolated topic, our framework is operationalised through a content-centered instructional model in which reflections on NOS/NOSI are intrinsically woven into experimental activities and historical discussions. This original combination addresses both the learning goals of a physics laboratory course and the broader epistemological goals of scientific literacy. By providing detailed experimental protocols alongside related NOS themes and facilitating student-led discussions, this study offers a rare operationalised bridge between authentic scientific practice and reflective historical instruction.

2. Research Design and Methodology

This study aims to investigate the effectiveness of an innovative instructional approach that combines authentic laboratory work with the history of science, with the intention of improving pre-service and trainee teachers’ understanding of the NOS and the NOSI among pre-service teachers and teacher students. Specifically, the research addresses the following questions:

• RQ1: To what extent does integrating subject matter learning and laboratory skills with historical reflections on Galileo’s experiments improve pre-service teachers’ understanding of the empirical and iterative nature of scientific knowledge, particularly in relation to the connection between experimental data and theoretical development?
• How does direct engagement in the modeling process, including stages of idealization and de-idealization, influence students’ perceptions of the role of models, inference, and the inherent nature of uncertainty in science? What views do students develop regarding the relationship between theories, laws, and models?
• RQ3: How do students interpret the processes of theory change and succession when confronted with discordant evidence? In what ways does deconstructing the ‘myth of the scientific method’—through a historical contrast between figures such as Galileo and Mersenne—foster a more nuanced understanding of the social and human dimensions of scientific practice, and encourage informed trust in science?
• RQ4: How does our historical–experimental approach facilitate the transition from a naïve, static view of science to an understanding of its provisional and cumulative nature, and what conceptual challenges remain?

The study sample consisted of 40 undergraduate students enrolled in a Physics Education laboratory course (Student teachers) and 25 pre-service physics teachers.

To address the research questions, a qualitative approach was adopted, focusing not on the quantitative measurement of knowledge, but on exploring the complexity of students’ epistemological structures and how these evolve following the intervention (Lising & Elby, 2005). In line with established Physics Education Research (PER) practices, we used a case study approach to analyze these phenomena in depth within an authentic classroom setting.

The focus is on processes of critical reflection and the conceptual tensions that emerge during the activities (Scherr & Hammer, 2009). Data collection primarily relied on written reflection protocols in the form of semi-structured written interviews, administered at key stages of the instructional sequence. These open-ended prompts were designed to stimulate critical thinking, capturing student feedback and metacognitive reflections (May & Etkina, 2002). Audio recordings supplemented these data, providing a granular account of the process and preserving the dynamics of group discussions (Conlin et al., 2007). Finally, we assessed disciplinary learning outcomes through traditional assessment tools and laboratory reports.

3. The Nature of Science and Scientific Inquiry: A Comprehensive Framework

The NOS and the NOSI constitute complementary and foundational frameworks for conceptualizing science as both a body of knowledge and a set of practices. While NOS addresses the epistemological foundations of science—emphasizing its empirical basis, tentative nature, and sociocultural embeddedness—NOSI highlights the procedural and methodological dimensions through which scientific knowledge is generated (Ladyman, 2012; Schwartz et al., 2004).

NOS is concerned with how scientific knowledge is developed, validated, and situated within broader philosophical, historical, and cultural contexts. Scientific knowledge is viewed not as an accumulation of objective facts but as an interpretative, creative, and socially mediated construction, grounded in empirical evidence and theory-laden observations (García-Carmona & Acevedo-Díaz, 2018; N. Lederman, 2007; N. G. Lederman, 1992). Observations are constrained by instruments, shaped by perceptual filters, and interpreted through theoretical frameworks, thus necessitating a conceptual distinction between observation and inference (N. G. Lederman et al., 2002). Consequently, science entails a mix of empirical, inferential, and socio-ethical dimensions, and its knowledge claims are provisional, open to revision, and historically situated (AAAS, 1990; Galili, 2019).

Scientific inquiry, as articulated in NOSI, is not a linear or algorithmic process but an iterative and self-correcting practice characterized by the formulation of investigable questions, experimental design, data analysis, and theory refinement (N. Lederman, 2007; Schwartz et al., 2012). Importantly, the distinction between NOS and NOSI remains contested. While Lederman (N. Lederman, 2007) argues for their conceptual separation, other scholars (e.g., Duschl, 2008; Crawford et al., 2013) maintain that they are intrinsically entangled, particularly through the mediating role of models in linking theory and data.

Modeling and idealization are central to both NOS and NOSI. Models serve as epistemic mediators between phenomena and theory (Morrison & Morgan, 1999), and almost invariably involve idealization—defined as the deliberate simplification or distortion of complex systems to render them analytically tractable (Matthews, 1994). Two forms of idealization can be distinguished: Aristotelian idealization, which involves the abstraction of irrelevant properties, and Galilean idealization, which introduces controlled distortions that allow for subsequent de-idealization. This distinction underscores the dynamic interplay between theory, model, and reality, and invites reflection on the conceptual status of scientific representations.

Scientific models can be categorized as data models—structured representations of experimental data—or theory models, which instantiate physical theories under specific assumptions and constraints. In both cases, the construction, testing, and revision of models are guided by inferential reasoning. Inference is thus a cornerstone of scientific knowledge production, encompassing processes such as hypothesis generation, prediction, and generalization from data. The terms prediction, forecast, and projection, while often used interchangeably, bear distinct meanings and implications for model-based reasoning (Keyfitz, 1972; Massad et al., 2005).

The predictive success of a theory is widely regarded as a critical criterion for its epistemic adequacy. Nonetheless, theories are understood to be provisional and fallible, even when well-supported by empirical evidence (N. G. Lederman et al., 2015). Within this framework, epistemic appraisal depends significantly on a theory’s falsifiability and its capacity to withstand rigorous attempts at refutation (Popper, 1959). Science is therefore both progressive and cumulative, building upon existing knowledge while remaining open to revision (Krauss, 2024a, 2024b, 2026). According to the Correspondence Principle, more general theories must reproduce the empirical successes of previous ones within their domain of validity (Popper, 1959). The coexistence of multiple theories within bounded validity domains (Chakravartty, 2017; Galili, 2013, 2019; Kuhn, 1962; S. Psillos, 2005) reflects the complexity and pluralism of scientific explanation.

Scientific reasoning also employs asymptotic reasoning and Gedankenexperimente (thought experiments) to explore model limits and discriminate between competing theories, often in the absence of direct empirical tests. The notion of an experimentum crucis—sometimes a conceptual rather than an empirical tool—helps us to understand how scientists make theory choices, which may be influenced by other criteria such as parsimony, explanatory coherence or heuristic power (Popper, 1959; Sober, 2008).

Recent scholarship has challenged the long-standing myth of a single, unified scientific method (Feyerabend, 1975; Kuhn, 1962), arguing instead that science employs a plurality of approaches, adapted to context and domain (Allchin, 2011; Osborne, 2014; Windschitl et al., 2008). Misrepresenting science as methodologically uniform risks obscuring its epistemic diversity and sociocultural situatedness.

Scientific knowledge is not produced in isolation but is subject to communal validation through mechanisms such as peer review, replication, and critical debate (Erduran, 2014; Höttecke & Allchin, 2020; Irzik & Nola, 2010). The historical case of Marin Mersenne illustrates the collective dimension of scientific practice, particularly in relation to communication and scrutiny of results. Similarly, Galileo’s work (Drake & MacLachlan, 1975; Schvartzer et al., 2020) reveals the theory-laden nature of observation and the conceptual role of idealization in experimentation.

The educational implications of these perspectives are profound. Developing students’ understanding of NOS and NOSI is essential for cultivating scientific literacy and a resilient, informed TIS—one that acknowledges its fallibility and capacity for self-correction.

Key concepts that emerge from this Comprehensive framework—and that will be further operationalized in the instructional sequence—include key topics reported in Table 1 and discussed in Appendix A.

Table 1. The key concepts in the Comprehensive framework (more details in Appendix A).

The Empirical Nature of Scientific Knowledge

2. Nature of Scientific Inquiry

3. Key Concepts in Scientific Inquiry, Models, and Theories

4. The Steps of Scientific Inquiry

5. Analysis of Collected Data to Draw Conclusions

6. Experimentation as a Process that Involves Modeling and Idealization

7. Modeling and Idealization

8. Aristotelian and Galilean Idealization

9. Models

10. Models of Data

11. Theory Model

12. Models as Mediators between a Theory and a Target System

13. Scientific Theories and Models: Nature, Role, and Representation in Education

14. Data, Observation, Inference, and Uncertainty

15. The Difference Between Observation and Inference

16. Statistical Inference

17. Inference as Prediction and Inference as the Construction of Scientific Theories

18. Uncertainty and the Probabilistic Nature of Scientific Knowledge: Uncertainty as Inherent to Scientific Knowledge

19. Aleatoric and Epistemic Uncertainty

20. Differences Between Prediction, Forecast, and Projection

21. Prediction and Evaluation of Theories

22. Predictive Power of Models and Theories and Falsification

23. Coexistence of Different Theories

24. Criteria for Choosing Between Competing Theories

25. Experimentum Crucis

26. Asymptotic Reasoning and Gedanken Experiment

27. Principle of Parsimony

28. Area/domain of Validity

29. Correspondence Principle

30. Nature of Scientific Knowledge

31. Tentativeness of Theories

32. The Self-Corrective NOS

33. The Progressive and Cumulative Character of Science

34. Inquiry as Iterative Process

35. Theory Ladenness and Creativity

36. Critique of the Myth of a Single Scientific Method

37. Social and Communal Aspects of Science

38. Social Embeddedness of Science

39. The Role of Communication and Scientific Community

40. Scientific Ethos

4. Instructional Design Methodology

The instructional design underlying our proposal is grounded in current research in science education, particularly regarding the teaching and learning of the NOS and the NOSI. Drawing upon an explicit and reflective pedagogy, the methodology integrates multiple interrelated components: guided inquiry-based instruction, hands-on experimentation, engaging activities, curricular integration, and the use of historical case studies.

A substantial body of research has emphasized the effectiveness of explicit and reflective approaches in fostering students’ understanding of NOS and NOSI (Abd-El-Khalick, 2011; Mesci & Schwartz, 2016; Cetin, 2021). In contrast to implicit or purely discovery-based models, explicit–reflective instruction makes the epistemic goals of learning activities visible and subject to discussion. Students are guided to actively construct conceptual understanding and subsequently engage in structured reflection, aimed at refining their views and comparing them with scientific norms (Vorholzer et al., 2018; N. G. Lederman et al., 2013). Reflection may occur during classroom dialogue, peer discussions, or written explanations, and it is crucial for connecting students’ experiences with broader epistemological principles. This approach resists the reduction of NOS to rote definitions, instead promoting conceptual depth through carefully scaffolded activities and critical engagement (Osborne, 2014).

Inquiry-Based Science Teaching (IBST) serves as a complementary framework for implementing explicit–reflective instruction. It is crucial to emphasize that the relevance of this approach extends beyond the ‘hard sciences’ to include the humanities, such as history and archeology, which are similarly grounded in rigorous inquiry and evidence-based methodology. Research suggests that engaging students in authentic inquiry—posing questions, planning investigations, collecting data, and justifying claims—can significantly enhance their grasp of NOSI dimensions (Capps & Crawford, 2013; Hrisa & Psillos, 2022). Effective inquiry involves not only procedural skills but also epistemic reflection, such as evaluating the limitations of evidence or recognizing the theory-ladenness of observations (Cetin, 2021). Such experiences allow students to practice science in ways that mirror actual scientific work, including the social and argumentative practices of the scientific community.

Another key dimension is the integration of NOS and NOSI into curricular content. Research shows that NOS instruction can be delivered both in content-embedded formats—within standard science topics such as physics, biology, or chemistry—or through decontextualized activities designed specifically to target epistemological understanding (N. Lederman & Abd-El-Khalick, 1998; Michel & Neumann, 2016; Schizas et al., 2022). The choice of context should be pedagogically motivated and aligned with the learning objectives of the unit, ensuring coherence between content and epistemic aims.

Lastly, our approach makes systematic use of historical case studies to illustrate the empirical and tentative nature of scientific knowledge. The history of science (HOS) offers a rich resource for highlighting the human dimension of science, including the influence of social, cultural, and political contexts on the development of scientific ideas (García-Carmona, 2018; Allchin, 2011). When carefully selected and accurately represented, historical narratives can reveal the processes by which scientific knowledge evolves, the role of controversy and consensus, and the significance of empirical evidence in theory change (N. Lederman, 2007). However, the effectiveness of HOS depends on its integration within an explicit–reflective framework to avoid misconceptions or oversimplification of scientific practice.

In sum, empirical studies suggest that explicit and reflective instruction, integrating inquiry-based learning and historical case analysis, is effective in fostering such understanding (Abd-El-Khalick, 2011; Mesci & Schwartz, 2016; Capps & Crawford, 2013). Experimental activities, such as replicating Galileo’s inclined plane investigations, can concretely illustrate the empirical NOS, the formulation of investigable questions, the construction and testing of models, and the interpretive nature of data. Thus, our instructional model draws on a convergence of pedagogical strategies that are empirically supported and epistemologically coherent. By combining Inquiry-Based Learning (IBL), explicit–reflective teaching, curricular integration, and the use of historical exemplars, the methodology seeks to cultivate a nuanced and authentic understanding of science as both a body of knowledge and a process of inquiry.

5. Scientific Inquiry and the Empirical Nature of Knowledge: A Case Study of Galileo’s Experiment

The learning sequence is presented in detail in Table 2, with each activity corresponding to one or more key concepts of the Comprehensive Framework.

Table 2. Detailed learning sequence, with each activity aligned to one or more key concepts of the Comprehensive Framework.

The Empirical Nature of Scientific Knowledge	The core activity of this educational module involves students reproducing Galileo’s experiment using a replica of his inclined plane apparatus. The primary objective is to measure the range of a ball launched from various heights on the plane and, crucially, to develop a model for the data. Students are encouraged to reflect on the fact that scientific knowledge is based on observation and real-world experience. This implies that scientific explanations must be supported by observational evidence and that answers to questions must be sought by interrogating nature. In other words, science does not rely on abstract hypotheses or theories but on data obtained through interaction with the natural world. Students are given minimal instructions, prompting them to partially design the experiment themselves.
Experimental Setup	The apparatus consists of an inclined plane (a rail) placed on a table (Figure 1). A ball is released from various heights (h) along the rail. Upon impact with the floor, a carbon paper sheet marks the landing position on a white sheet. The range (X) is measured as the horizontal distance from the point directly below the rail’s edge to the impact mark.
Nature of Scientific Inquiry Formulating questions that can be empirically investigated. Designing experiments to test hypotheses	Steps of scientific inquiry: Formulating empirically investigable questions. Designing experiments to test hypotheses. Collecting and analyzing data. The activity is rooted in gathering and analyzing empirical data from observations and experiments. The validity and reliability of the developed models are continually assessed by comparing their predictions with experimental results. This highlights how scientific knowledge derives from and is tested against observations of the natural world. A key aspect of this activity, aligned with the NOSI, is its guidance by an explicit question posed to students: “Is it possible to find a relationship between the ball’s range and the release height?”
Uncertainty and the Probabilistic Nature of Scientific Knowledge Uncertainty as inherent to scientific knowledge Analysis of collected data to draw conclusions	To account for variability, multiple measurements are taken for each release height, and the average value is recorded (Figure 2). The uncertainty in range measurements (dispersion of impact points) is considered alongside other measurement uncertainties (e.g., launch height, table height). The experiment provides a concrete understanding of how aleatoric (random) uncertainty affects range estimates, allowing students to visualize measurement dispersion. Subsequently, students analyze the data using statistical tools and dedicated software, producing graphical representations of results. At this stage, students compare their findings with hypotheses about how range depends on launch height.
Experimentation as a Modeling Process The models of data	After data collection, students engage in modeling and analysis. They construct a data model based on experimental results (Figure 3). “The models students develop are grounded in their intuitions, the availability of software tools, and considerations of uncertainty.” Typically, students propose two models: (1) A linear model. (2) A quadratic polynomial model. Interestingly, over 50% of students use a quadratic curve. Instead of correcting them immediately, this serves as an opportunity to discuss the nature of data models. Students are also encouraged to incorporate uncertainty by propagating interpolation parameter errors.
Testing models Inference as prediction Predictive power of models and theories	The model must be tested for reliability. Students are asked to estimate, within a margin of uncertainty, the range corresponding to an intermediate height using statistical inference (Figure 4). To evaluate predictive power, this estimate is compared with the actual range measured in ten additional trials. The results show that predictions from both linear and polynomial models align with experimental data within uncertainty margins, demonstrating their compatibility. From an educational and NOS perspective, this phase is critical because it allows students to: Develop hypotheses. Analyze data. Draw conclusions. Verify the consistency between experimental data and hypotheses. It also clarifies the distinction between observations and inferences. Here, inference has a dual meaning: Statistical inference: Quantitative predictions based on data analysis. Theoretical inference: Deriving scientifically meaningful models from data. The concept of predictive uncertainty is emphasized, reflecting the reliability of estimates due to statistical noise. Students also explore key modeling concepts like prediction, forecast, and projection, underscoring science’s role in interpreting and predicting phenomena.
Group Comparison Communication of results for review and validation. Theory Laden (paradigm guiding model construction) and Creativity. Epistemic uncertainty; influence of methods on outcomes.	Students are encouraged to compare results to stimulate discussion, experiencing the peer review’s role in scientific knowledge development. “Divergent results highlight the coexistence of different theories and epistemic uncertainty between models.” The so-called replication crisis is introduced, where identical procedures may yield different outcomes due to methodological subjectivity. This underscores how methods influence results and the role of subjectivity in scientific knowledge construction. Epistemic uncertainty, which produces a systematic uncertainty, arises when multiple models explain the same data. Different methods—in data collection, analysis, or interpretation—significantly affect outcomes. Model choice reflects theoretical assumptions, highlighting the subjectivity inherent in scientific inquiry.
Model Testing and Falsification Area of validity	At this stage of the work, students are asked to compare the theoretical prediction obtained from the model with that relating to a range corresponding to a height outside the data interval used previously (Figure 5). This step explicitly introduces the concept of the validity range of a model or theory, emphasizing how, in principle, each model should only be considered reliable within the domain of the experimental data from which it was derived (Figure 6). The main objective is to test the behavior of the model beyond its initial range, examining whether it can be extended or whether, on the contrary, it encounters significant limitations.
Asymptotic reasoning and Gedanken Experiment	To do this, two fundamental conceptual tools are introduced: asymptotic reasoning and thought experiments (Gedankenexperiment). These tools make it possible to evaluate the validity of a model even in the absence of real experiments, allowing, in some cases, to refute it logically. At this point, a new guiding question is formulated: Can the model be considered reliable even outside the initial height range, or does it show structural limitations? To answer this question, two ideal but significant situations are analyzed (Figure 5), which allow the models to be tested through limit reasoning: Behavior at infinity: if we imagine increasing the release height of the ball indefinitely, we would expect the range to also increase without limits. However, not all models comply with this expected behavior. Behavior at zero height: if the ball is released from zero height, it is reasonable to expect it to fall vertically and its range to be zero. Some models, however, predict a finite range. Through these analyses, even in the absence of direct experimental verification, it is possible to identify inconsistencies in the model and arrive at its logical falsification.
Experimentum crucis Provisional and conditional nature of scientific theories	This also introduces the concept of experimentum crucis, or a decisive test between competing models, which is useful for discriminating which one best describes reality. This exercise leads us to reflect on the provisional and conditional nature of scientific theories, showing that the predictive value of a model depends strictly on its domain of validity, and that any extension must be justified, both conceptually and empirically.
Model Revision Constructing a new data model	Students are then invited to test a power-law regression (initially proposed by only 20%). The optimal fit yields an exponent close to 0.5, producing a monotonically increasing curve with a zero-intercept—consistent with asymptotic expectations.
Test the model using the old data: Correspondence principle Criteria for Choosing Between Competing Theories	Once a new model that respects asymptotic behavior is constructed, its validity is checked within the old data range. According to the correspondence principle, a new model should not only fit new data but also reproduce verified predictions of previous theories within their validity domain as a limiting case. This means that in the region where previous models were not refuted by data, the new and old models can coexist. In such situations, when data does not allow falsification, the choice between models is not based on empirical criteria but on epistemic or aesthetic ones, such as preferring the simplest (parsimony) or most theoretically coherent model.
Theoretical Models and Idealization The idealizations in theoretical models Models as mediators between a theory and a target system De-idealization	Students are asked to construct a theoretical model from Newtonian mechanics to explain the experimental data (Figure 7). The initial tendency is to use a simplified model (point mass on a frictionless plane), a strong Galilean idealization. While simple, this model (predicting R(h) = √4gh) significantly overestimates the range compared to experimental data. This discrepancy leads students to reflect on the idealizations in their theoretical model. They discover that theory and data are not compared directly, but rather two models: a theoretical, idealized model derived from theory, and an experimental model representing data with uncertainty. Scientific models act as crucial mediators between abstract theories and the empirical world…. Faced with the discrepancy, a process of de-idealization begins to make the model more realistic. Considering the marble as a rigid sphere rolling without slipping, introducing rotational energy, gives a formula (R(h) = √(10/7)gh) closer but still not entirely accurate. Further idealizations, like the marble rolling on a rail (implying a smaller effective radius), lead to an updated model. The activity shows that a perfect representation of reality is impossible; every model remains a simplification.
Iterative construction of scientific knowledge	Students experience the importance of modeling, idealization limits, and continuous revision in the construction of scientific knowledge.
Tentativeness of theories	During the course, students were able to experience first-hand how scientific theories—or the models derived from them—are by their very nature provisional and subject to revision. The emergence of new experimental evidence can call their validity into question, leading to their modification, replacement or extension. This provisional nature is not a sign of weakness, but an essential feature of science: it advances cumulatively and gradually, progressively integrating new data and improving its explanations. This dynamic does not imply that science provides incorrect answers, but rather that its results are always open to improvement and tend to offer explanations with the lowest possible degree of uncertainty and the highest level of reliability. The evolution of theories is often driven by the reduction of approximations, the replacement of no longer sustainable idealizations, and the integration of new observed phenomena, without losing the consolidated results in their domains of validity. In this context, students are asked to critically reflect on the meaning of the provisional nature of scientific knowledge: how much it affects the development of knowledge and what educational value it has. Through the practice of comparing models, revising initial hypotheses, and analyzing discrepancies between theory and data, students experience first-hand that uncertainty and change do not undermine the reliability of science, but rather constitute its lifeblood. The aim of this reflection is to stimulate a conception of science not as a static set of absolute truths, but as a constantly evolving process based on critical dialogue, experimental verification and the collective construction of knowledge.
Inquiry as iterative process	We therefore invite students to reflect on the path they have taken and how knowledge has been constructed: an iterative process, made up of attempts, adjustments and continuous comparisons between theoretical models and experimental evidence. Despite the effort put into refining the models, no perfect agreement has been reached between theory and experiment—and this is precisely the key point.
The Self-Correcting nature of science The Progressive and Cumulative Character of Science	In fact, the journey highlights one of the fundamental aspects of science: its self-correcting nature. Far from being a rigid and infallible system, scientific knowledge evolves precisely through the analysis of discrepancies, the revision of hypotheses and the expansion of models. Every step forward, even if partial or provisional, contributes to the progress of knowledge and its cumulative development. Students are therefore invited to discuss this aspect, asking themselves what it really means to ‘understand’ in the scientific field: not so much arriving at a definitive truth, but learning to construct increasingly reliable and coherent models, while accepting the limitations and uncertainties that inevitably accompany every stage of research. In this way, a more realistic, mature and critical view of science is promoted, as an open, dynamic process based on the continuous comparison between theory and experience.
Historical Reflection: Galileo’s Experiment and Drake’s Reconstruction
	At this point, we propose a retrospective reinterpretation of the entire experiment in light of Galileo’s historic experiment, as analyzed by Drake and MacLachlan (Drake & MacLachlan, 1975) through the famous folio 116v. The students’ explanations tend to combine experimental data with prior knowledge, but this does not seem to have been the case for Galileo. Drake shows how the results reported in the folio reveal a clear discrepancy between the experimental data and the theoretical data calculated by Galileo.
Theory Laden (paradigm guiding model construction) and Creativity	Galileo seems, however, to correctly interpret the parabolic motion of the projectile, resulting from the combination of a uniformly accelerated motion along the vertical (whose ‘ratio’ is geometric, i.e., quadratic with respect to time) and a uniform rectilinear motion along the horizontal (with an arithmetic ‘ratio’, i.e., linear in time). It is clear that Galileo already had expectations about the results of the experiment, that he gave priority to predictions based on his intuition, and that his theoretical intuition guided him in conducting and interpreting the experiment. Students are invited to reflect on the role of creativity in the development of scientific knowledge and on the relationship between theoretical intuition and experimentation.
Critique of the myth of a single scientific method	The annotation ‘should be’, with which Galileo accompanies the data on folio 116, suggests that he believed the experiment to be affected by some kind of error—not clearly identified—but that this did not lead him to question the theory. Although it was not customary at the time to use experiments to prove theories, this methodological choice appears to be in stark contrast to the often idealized image of Galileo as an unyielding opponent of dogmatism and the founding father of the scientific method. In reality, Galileo behaved pragmatically, showing a willingness to set aside the experiment in order to safeguard a theory he believed to be coherent and well-founded. This episode calls into question a simplified and widespread narrative, also conveyed by numerous school textbooks, which presents a mythologized figure of Galileo and a conception of the scientific method reduced to a rigid sequence of steps—almost a textbook recipe. Students are presented with this mythological view of the scientific method, according to which Galileo, faced with data contrary to his predictions, was forced to modify his theory. This representation is contrasted with Drake’s interpretation, according to which Galileo chose instead to ignore the conflicting data in order to keep his theory intact. At the same time, we present students with Feyerabend’s reflection, who criticized Galileo Galilei’s scientific method, arguing that the spread of his theory was not due to its intrinsic logic, but to his propaganda. According to Feyerabend, Galileo used communication strategies to persuade, rather than relying on a rigorous scientific method. The history of science seems to reveal a different methodology in the construction of scientific knowledge. At this point, students are invited to reflect: what do they think of a ‘fraudulent’ Galileo who deliberately ignores contrary evidence? And, more generally, what is their idea of the scientific method?
Historical Reflection: Mersenne as the founder of the scientific community
The role of Communication and Scientific Community Communication of results for review and validation Scientific community	We introduced students to the historical figure of Marin Mersenne, who can be considered one of the main architects of the birth of the European scientific community. Mersenne was a tireless coordinator of experiments, a rigorous critic and a promoter of the circulation of knowledge. His vast network of correspondence, active well before the birth of scientific journals, constituted a veritable infrastructure for communication between European thinkers. In this sense, he anticipated and embodied the modern ideal of collaborative and public science. Mersenne was also an extraordinary cultural mediator: he introduced Galilean mechanics to France. He also hosted a scientific salon in his Parisian home where figures such as Descartes, Hobbes, Fermat, Huygens and Torricelli gathered. Thanks to him, a culture of science developed based on the continuous exchange of ideas, critical discussion and the importance of communication.
The role of peer review and scientific debate in the self-correcting nature of science	Mersenne is often remembered as the first true ‘reviewer’ of Galileo’s works, particularly for his experimental verification during his trip to Italy in 1644–1645, when he met Torricelli. His position was not that of a mere philosophical opponent, but of a convinced empiricist, determined to subject Galileo’s claims to rigorous scrutiny, including those on the isochronism of the pendulum and the motion of bodies along an inclined plane. Mersenne demonstrated that large oscillations take longer than small ones, contrary to what Galileo claimed through the character of Salviati in Dialogue Concerning the Two Chief World Systems. Today we know that Galileo was wrong: isochronism is only a valid approximation for small oscillations, typical of simple harmonic motion.
	This episode shows that Galileo was not immune to criticism and, indeed, that his methodology was deeply ‘theory-laden’, i.e., guided by theoretical premises rather than neutral observations. In this context, the figure of Mersenne challenges the mythologized narrative of Galileo as the sole father of the scientific method, revealing how much the construction of knowledge is a collective process, open to discussion and peer review.
Scientific ethos; peer review as a driver of correction	This historical reflection also provides an opportunity to address the issue of the ethics of science, understood both as the morality of ends (i.e., the purpose of research) and of means (i.e., the methods used). Students are asked whether they believe that the scientific community acts morally and how they perceive its immorality, if any. The comparison between Galileo and Mersenne can serve as an emblematic case: on the one hand, the tension between theory and data; on the other, the emergence of a collective and self-critical scientific method, in which dissent and peer review are fundamental to the advancement of knowledge.

Figure 1. (Left) The setup: the inclined plane is a pre-built rail on which marbles can roll. This is put on a table in such a way that marbles leave the rail and fall off the table, reaching the floor. The marbles hit a carbon paper sheet put on a white sheet, so that the position of landing is recorded. (Right) Students and teachers during a laboratory activity, and the results of their data interpolation.

Figure 2. Students then repeat the experiment at least ten times for each chosen height. In this way, they can estimate the range as the average of the ten results and the error as the semi-dispersion (or the variance if the repetitions are enough. (Left) A photo of the position marks on the paper and the corresponding histograms. (Right) The experimental points of the range versus launch height.

Figure 3. (Left) The linear model. (Right) A quadratic polynomial. The shaded regions represent uncertainties.

Figure 4. Comparison of the models with experimental data and test of the models with a new experimental point.

Figure 5. Representation of asymptotic limits and model predictions.

Figure 6. Area of validity and coexistence of different models.

Figure 7. Experimental data and model of data compared with different models of theory. Theoretical curve obtained with the point mass and frictionless plane model, compared with the power law model of data. Theoretical curve obtained with the rolling marble on a plane model. Theoretical curve obtained with the rolling ball on a rail model. Schematic drawing of the marble in the rail, showing the effective radius for the rolling motion.

Table 2. Detailed learning sequence, with each activity aligned to one or more key concepts of the Comprehensive Framework.

The Empirical Nature of Scientific Knowledge	The core activity of this educational module involves students reproducing Galileo’s experiment using a replica of his inclined plane apparatus. The primary objective is to measure the range of a ball launched from various heights on the plane and, crucially, to develop a model for the data. Students are encouraged to reflect on the fact that scientific knowledge is based on observation and real-world experience. This implies that scientific explanations must be supported by observational evidence and that answers to questions must be sought by interrogating nature. In other words, science does not rely on abstract hypotheses or theories but on data obtained through interaction with the natural world. Students are given minimal instructions, prompting them to partially design the experiment themselves.
Experimental Setup	The apparatus consists of an inclined plane (a rail) placed on a table (Figure 1). A ball is released from various heights (h) along the rail. Upon impact with the floor, a carbon paper sheet marks the landing position on a white sheet. The range (X) is measured as the horizontal distance from the point directly below the rail’s edge to the impact mark.
Nature of Scientific Inquiry Formulating questions that can be empirically investigated. Designing experiments to test hypotheses	Steps of scientific inquiry: Formulating empirically investigable questions. Designing experiments to test hypotheses. Collecting and analyzing data. The activity is rooted in gathering and analyzing empirical data from observations and experiments. The validity and reliability of the developed models are continually assessed by comparing their predictions with experimental results. This highlights how scientific knowledge derives from and is tested against observations of the natural world. A key aspect of this activity, aligned with the NOSI, is its guidance by an explicit question posed to students: “Is it possible to find a relationship between the ball’s range and the release height?”
Uncertainty and the Probabilistic Nature of Scientific Knowledge Uncertainty as inherent to scientific knowledge Analysis of collected data to draw conclusions	To account for variability, multiple measurements are taken for each release height, and the average value is recorded (Figure 2). The uncertainty in range measurements (dispersion of impact points) is considered alongside other measurement uncertainties (e.g., launch height, table height). The experiment provides a concrete understanding of how aleatoric (random) uncertainty affects range estimates, allowing students to visualize measurement dispersion. Subsequently, students analyze the data using statistical tools and dedicated software, producing graphical representations of results. At this stage, students compare their findings with hypotheses about how range depends on launch height.
Experimentation as a Modeling Process The models of data	After data collection, students engage in modeling and analysis. They construct a data model based on experimental results (Figure 3). “The models students develop are grounded in their intuitions, the availability of software tools, and considerations of uncertainty.” Typically, students propose two models: (1) A linear model. (2) A quadratic polynomial model. Interestingly, over 50% of students use a quadratic curve. Instead of correcting them immediately, this serves as an opportunity to discuss the nature of data models. Students are also encouraged to incorporate uncertainty by propagating interpolation parameter errors.
Testing models Inference as prediction Predictive power of models and theories	The model must be tested for reliability. Students are asked to estimate, within a margin of uncertainty, the range corresponding to an intermediate height using statistical inference (Figure 4). To evaluate predictive power, this estimate is compared with the actual range measured in ten additional trials. The results show that predictions from both linear and polynomial models align with experimental data within uncertainty margins, demonstrating their compatibility. From an educational and NOS perspective, this phase is critical because it allows students to: Develop hypotheses. Analyze data. Draw conclusions. Verify the consistency between experimental data and hypotheses. It also clarifies the distinction between observations and inferences. Here, inference has a dual meaning: Statistical inference: Quantitative predictions based on data analysis. Theoretical inference: Deriving scientifically meaningful models from data. The concept of predictive uncertainty is emphasized, reflecting the reliability of estimates due to statistical noise. Students also explore key modeling concepts like prediction, forecast, and projection, underscoring science’s role in interpreting and predicting phenomena.
Group Comparison Communication of results for review and validation. Theory Laden (paradigm guiding model construction) and Creativity. Epistemic uncertainty; influence of methods on outcomes.	Students are encouraged to compare results to stimulate discussion, experiencing the peer review’s role in scientific knowledge development. “Divergent results highlight the coexistence of different theories and epistemic uncertainty between models.” The so-called replication crisis is introduced, where identical procedures may yield different outcomes due to methodological subjectivity. This underscores how methods influence results and the role of subjectivity in scientific knowledge construction. Epistemic uncertainty, which produces a systematic uncertainty, arises when multiple models explain the same data. Different methods—in data collection, analysis, or interpretation—significantly affect outcomes. Model choice reflects theoretical assumptions, highlighting the subjectivity inherent in scientific inquiry.
Model Testing and Falsification Area of validity	At this stage of the work, students are asked to compare the theoretical prediction obtained from the model with that relating to a range corresponding to a height outside the data interval used previously (Figure 5). This step explicitly introduces the concept of the validity range of a model or theory, emphasizing how, in principle, each model should only be considered reliable within the domain of the experimental data from which it was derived (Figure 6). The main objective is to test the behavior of the model beyond its initial range, examining whether it can be extended or whether, on the contrary, it encounters significant limitations.
Asymptotic reasoning and Gedanken Experiment	To do this, two fundamental conceptual tools are introduced: asymptotic reasoning and thought experiments (Gedankenexperiment). These tools make it possible to evaluate the validity of a model even in the absence of real experiments, allowing, in some cases, to refute it logically. At this point, a new guiding question is formulated: Can the model be considered reliable even outside the initial height range, or does it show structural limitations? To answer this question, two ideal but significant situations are analyzed (Figure 5), which allow the models to be tested through limit reasoning: Behavior at infinity: if we imagine increasing the release height of the ball indefinitely, we would expect the range to also increase without limits. However, not all models comply with this expected behavior. Behavior at zero height: if the ball is released from zero height, it is reasonable to expect it to fall vertically and its range to be zero. Some models, however, predict a finite range. Through these analyses, even in the absence of direct experimental verification, it is possible to identify inconsistencies in the model and arrive at its logical falsification.
Experimentum crucis Provisional and conditional nature of scientific theories	This also introduces the concept of experimentum crucis, or a decisive test between competing models, which is useful for discriminating which one best describes reality. This exercise leads us to reflect on the provisional and conditional nature of scientific theories, showing that the predictive value of a model depends strictly on its domain of validity, and that any extension must be justified, both conceptually and empirically.
Model Revision Constructing a new data model	Students are then invited to test a power-law regression (initially proposed by only 20%). The optimal fit yields an exponent close to 0.5, producing a monotonically increasing curve with a zero-intercept—consistent with asymptotic expectations.
Test the model using the old data: Correspondence principle Criteria for Choosing Between Competing Theories	Once a new model that respects asymptotic behavior is constructed, its validity is checked within the old data range. According to the correspondence principle, a new model should not only fit new data but also reproduce verified predictions of previous theories within their validity domain as a limiting case. This means that in the region where previous models were not refuted by data, the new and old models can coexist. In such situations, when data does not allow falsification, the choice between models is not based on empirical criteria but on epistemic or aesthetic ones, such as preferring the simplest (parsimony) or most theoretically coherent model.
Theoretical Models and Idealization The idealizations in theoretical models Models as mediators between a theory and a target system De-idealization	Students are asked to construct a theoretical model from Newtonian mechanics to explain the experimental data (Figure 7). The initial tendency is to use a simplified model (point mass on a frictionless plane), a strong Galilean idealization. While simple, this model (predicting R(h) = √4gh) significantly overestimates the range compared to experimental data. This discrepancy leads students to reflect on the idealizations in their theoretical model. They discover that theory and data are not compared directly, but rather two models: a theoretical, idealized model derived from theory, and an experimental model representing data with uncertainty. Scientific models act as crucial mediators between abstract theories and the empirical world…. Faced with the discrepancy, a process of de-idealization begins to make the model more realistic. Considering the marble as a rigid sphere rolling without slipping, introducing rotational energy, gives a formula (R(h) = √(10/7)gh) closer but still not entirely accurate. Further idealizations, like the marble rolling on a rail (implying a smaller effective radius), lead to an updated model. The activity shows that a perfect representation of reality is impossible; every model remains a simplification.
Iterative construction of scientific knowledge	Students experience the importance of modeling, idealization limits, and continuous revision in the construction of scientific knowledge.
Tentativeness of theories	During the course, students were able to experience first-hand how scientific theories—or the models derived from them—are by their very nature provisional and subject to revision. The emergence of new experimental evidence can call their validity into question, leading to their modification, replacement or extension. This provisional nature is not a sign of weakness, but an essential feature of science: it advances cumulatively and gradually, progressively integrating new data and improving its explanations. This dynamic does not imply that science provides incorrect answers, but rather that its results are always open to improvement and tend to offer explanations with the lowest possible degree of uncertainty and the highest level of reliability. The evolution of theories is often driven by the reduction of approximations, the replacement of no longer sustainable idealizations, and the integration of new observed phenomena, without losing the consolidated results in their domains of validity. In this context, students are asked to critically reflect on the meaning of the provisional nature of scientific knowledge: how much it affects the development of knowledge and what educational value it has. Through the practice of comparing models, revising initial hypotheses, and analyzing discrepancies between theory and data, students experience first-hand that uncertainty and change do not undermine the reliability of science, but rather constitute its lifeblood. The aim of this reflection is to stimulate a conception of science not as a static set of absolute truths, but as a constantly evolving process based on critical dialogue, experimental verification and the collective construction of knowledge.
Inquiry as iterative process	We therefore invite students to reflect on the path they have taken and how knowledge has been constructed: an iterative process, made up of attempts, adjustments and continuous comparisons between theoretical models and experimental evidence. Despite the effort put into refining the models, no perfect agreement has been reached between theory and experiment—and this is precisely the key point.
The Self-Correcting nature of science The Progressive and Cumulative Character of Science	In fact, the journey highlights one of the fundamental aspects of science: its self-correcting nature. Far from being a rigid and infallible system, scientific knowledge evolves precisely through the analysis of discrepancies, the revision of hypotheses and the expansion of models. Every step forward, even if partial or provisional, contributes to the progress of knowledge and its cumulative development. Students are therefore invited to discuss this aspect, asking themselves what it really means to ‘understand’ in the scientific field: not so much arriving at a definitive truth, but learning to construct increasingly reliable and coherent models, while accepting the limitations and uncertainties that inevitably accompany every stage of research. In this way, a more realistic, mature and critical view of science is promoted, as an open, dynamic process based on the continuous comparison between theory and experience.
Historical Reflection: Galileo’s Experiment and Drake’s Reconstruction
	At this point, we propose a retrospective reinterpretation of the entire experiment in light of Galileo’s historic experiment, as analyzed by Drake and MacLachlan (Drake & MacLachlan, 1975) through the famous folio 116v. The students’ explanations tend to combine experimental data with prior knowledge, but this does not seem to have been the case for Galileo. Drake shows how the results reported in the folio reveal a clear discrepancy between the experimental data and the theoretical data calculated by Galileo.
Theory Laden (paradigm guiding model construction) and Creativity	Galileo seems, however, to correctly interpret the parabolic motion of the projectile, resulting from the combination of a uniformly accelerated motion along the vertical (whose ‘ratio’ is geometric, i.e., quadratic with respect to time) and a uniform rectilinear motion along the horizontal (with an arithmetic ‘ratio’, i.e., linear in time). It is clear that Galileo already had expectations about the results of the experiment, that he gave priority to predictions based on his intuition, and that his theoretical intuition guided him in conducting and interpreting the experiment. Students are invited to reflect on the role of creativity in the development of scientific knowledge and on the relationship between theoretical intuition and experimentation.
Critique of the myth of a single scientific method	The annotation ‘should be’, with which Galileo accompanies the data on folio 116, suggests that he believed the experiment to be affected by some kind of error—not clearly identified—but that this did not lead him to question the theory. Although it was not customary at the time to use experiments to prove theories, this methodological choice appears to be in stark contrast to the often idealized image of Galileo as an unyielding opponent of dogmatism and the founding father of the scientific method. In reality, Galileo behaved pragmatically, showing a willingness to set aside the experiment in order to safeguard a theory he believed to be coherent and well-founded. This episode calls into question a simplified and widespread narrative, also conveyed by numerous school textbooks, which presents a mythologized figure of Galileo and a conception of the scientific method reduced to a rigid sequence of steps—almost a textbook recipe. Students are presented with this mythological view of the scientific method, according to which Galileo, faced with data contrary to his predictions, was forced to modify his theory. This representation is contrasted with Drake’s interpretation, according to which Galileo chose instead to ignore the conflicting data in order to keep his theory intact. At the same time, we present students with Feyerabend’s reflection, who criticized Galileo Galilei’s scientific method, arguing that the spread of his theory was not due to its intrinsic logic, but to his propaganda. According to Feyerabend, Galileo used communication strategies to persuade, rather than relying on a rigorous scientific method. The history of science seems to reveal a different methodology in the construction of scientific knowledge. At this point, students are invited to reflect: what do they think of a ‘fraudulent’ Galileo who deliberately ignores contrary evidence? And, more generally, what is their idea of the scientific method?
Historical Reflection: Mersenne as the founder of the scientific community
The role of Communication and Scientific Community Communication of results for review and validation Scientific community	We introduced students to the historical figure of Marin Mersenne, who can be considered one of the main architects of the birth of the European scientific community. Mersenne was a tireless coordinator of experiments, a rigorous critic and a promoter of the circulation of knowledge. His vast network of correspondence, active well before the birth of scientific journals, constituted a veritable infrastructure for communication between European thinkers. In this sense, he anticipated and embodied the modern ideal of collaborative and public science. Mersenne was also an extraordinary cultural mediator: he introduced Galilean mechanics to France. He also hosted a scientific salon in his Parisian home where figures such as Descartes, Hobbes, Fermat, Huygens and Torricelli gathered. Thanks to him, a culture of science developed based on the continuous exchange of ideas, critical discussion and the importance of communication.
The role of peer review and scientific debate in the self-correcting nature of science	Mersenne is often remembered as the first true ‘reviewer’ of Galileo’s works, particularly for his experimental verification during his trip to Italy in 1644–1645, when he met Torricelli. His position was not that of a mere philosophical opponent, but of a convinced empiricist, determined to subject Galileo’s claims to rigorous scrutiny, including those on the isochronism of the pendulum and the motion of bodies along an inclined plane. Mersenne demonstrated that large oscillations take longer than small ones, contrary to what Galileo claimed through the character of Salviati in Dialogue Concerning the Two Chief World Systems. Today we know that Galileo was wrong: isochronism is only a valid approximation for small oscillations, typical of simple harmonic motion.
	This episode shows that Galileo was not immune to criticism and, indeed, that his methodology was deeply ‘theory-laden’, i.e., guided by theoretical premises rather than neutral observations. In this context, the figure of Mersenne challenges the mythologized narrative of Galileo as the sole father of the scientific method, revealing how much the construction of knowledge is a collective process, open to discussion and peer review.
Scientific ethos; peer review as a driver of correction	This historical reflection also provides an opportunity to address the issue of the ethics of science, understood both as the morality of ends (i.e., the purpose of research) and of means (i.e., the methods used). Students are asked whether they believe that the scientific community acts morally and how they perceive its immorality, if any. The comparison between Galileo and Mersenne can serve as an emblematic case: on the one hand, the tension between theory and data; on the other, the emergence of a collective and self-critical scientific method, in which dissent and peer review are fundamental to the advancement of knowledge.

6. Results

The reflections collected from students at the conclusion of the instructional sequence present a rich and complex picture. Students demonstrate both promising conceptual shifts and persistent epistemological tensions regarding the NOS and the NOSI. These responses offer valuable insights into how learners engage with epistemological themes when these are made explicit in teaching.

One salient theme pertains to the empirical basis of scientific knowledge. Students generally acknowledge the foundational role of observation and experimentation in science. However, ambiguities persist regarding how new scientific knowledge is generated. While all students agreed with the statement that “scientific knowledge is constructed through a process that is based on observations of nature,” a significant portion expressed views reflecting a linear or static model of knowledge acquisition. Approximately 25% endorsed a realist conception, stating that “Nature follows exact laws and scientists discover these laws,” and about 33% believed that “new scientific knowledge derives mainly from the development of existing scientific theories.” These responses suggest a partial internalization of NOS principles, with some students oscillating between empiricist and rationalist paradigms.

Further epistemological insights were elicited through students’ reactions to a passage from Galileo’s Il Saggiatore, in which nature is described as a book written in mathematical language. One student countered this view, stating: “It is the opposite, it is mathematics that is constructed by taking inspiration from nature.” This position reflects a descriptive rather than constitutive view of mathematics in science, opposing Galileo’s Platonist stance. Other students highlighted how exposure to NOS education shifts their perspectives over time, noting that early beliefs about the perfect descriptive power of mathematics evolved into more critical stances after engagement with scientific practice.

The role of falsification emerged prominently. Many students expressed that theories should be revised when confronted with contradictory data. However, most favored a cumulative model of theory change rather than wholesale replacement. One student remarked, “A single result is not enough to determine that the new theory is better than the old one... numerous experimental data are needed.” Others warned of the risk of ad hoc modifications that preserve a theory beyond its rational limits: “You risk adding pieces until they are no longer pieces of the theory and then you have to start from scratch.”

Students also articulated nuanced views on predictive capacity and theory choice. Elegance and formalism were seen as secondary to empirical adequacy. One student wrote, “Although formalism and elegance can be criteria, it is the ability to predict experimental results that matters most.” Importantly, many referenced the correspondence principle, noting that new theories should reproduce the successful predictions of the older ones within their valid domain. This reflects a sophisticated understanding of theory succession and scientific continuity.

The concept of uncertainty and tentativeness in science showed substantial development. Initially, most students equated science with absolute truths; post-instruction, this figure dropped to about 25%. Still, some students misunderstood uncertainty as a lack of reliability. Others, however, offered insightful statements: “Uncertainty is not a defect, but what makes science honest.” The tentativeness of scientific knowledge was linked to the dynamic and self-corrective nature of the scientific enterprise.

Discussions around scientific theories, laws, and models revealed considerable conceptual differences. While students generally understood that laws and theories are distinct, their definitions diverged. Some viewed theories as higher-order constructs encompassing laws, while others reversed this hierarchy. Different views were also expressed regarding whether models precede or follow theories. One student claimed, “First comes the model and then the theory,” while another suggested, “A theory includes different models.” Despite the lack of consensus, most students agreed that models are idealized representations and approximations rather than exact mirrors of reality. Models were also recognized as mediators between theory and phenomena. Students noted that theories cannot be directly tested, but only through models: “Experimentally I cannot directly verify a theory; I have to go to the model that represents reality.” Some emphasized that models assist in prediction and explanation, while others highlighted the role of models in theory formation. As one student noted, “I take the data, develop a model, look for a theory, and compare it with new data.” Regarding scientific laws, students’ views were similarly diverse. Some asserted that laws are embedded within theories; others argued that laws are descriptive generalizations independent of theoretical frameworks. These inconsistencies indicate the presence of divergent opinions regarding the hierarchy and interrelation of scientific constructs. Notably, students proposed informal hierarchies: “The theory is something bigger than the law,” or “The model is not necessarily true, while laws and theories are.” Some even viewed theory as the endpoint of inquiry: “For me, theory is the arrival, the ultimate conclusion of the knowledge process.”

The scientific method was initially perceived as a rigid, algorithmic process. One student described it as “the set of rules for distinguishing scientific things from opinions.” However, through discussion and engagement, students came to appreciate a more flexible and creative view. As one put it, “Science should not be understood as a monotonous and uniform process, but as a set of adaptable approaches.”

Students also acknowledged the role of imagination and creativity in science. The idea that creativity and rationality coexist was widely accepted. One noted, “Imagination allows us to imagine paradoxes that challenge our theories.” Theoretical bias (theory-ladenness) was also acknowledged, with several students stating that “experiments are always inserted in a theoretical context.”

Students consistently emphasized the human and subjective dimension of science. Reflections on the replication crisis led to recognition that data interpretation and model selection are inherently influenced by personal and contextual factors. Statements like “There is always the intuition of the single person” and “Imagination is important for interpreting data” underscore this view.

Students also recognized the pluralism and coexistence of models and theories. One student stated, “Certain phenomena have a dual nature that can be explained by different models, until a broader theory is developed.” Others emphasized the domain-specific validity of models, suggesting a practical rather than absolutist epistemology.

The social dimension of science was another emergent theme. Students understood that scientific progress is collective, not individual. One commented, “Often the debate is presented as a one-on-one between two scientists, when in reality it is at the community level.” The self-correcting NOS was linked to institutional practices such as peer review and reproducibility.

External influences on science were acknowledged as well. Nearly half of the students mentioned factors like funding, personal bias, and historical context as shaping scientific outcomes. Nevertheless, TIS remained strong, grounded in its operational success and capacity for revision. One student noted, “Science is the best tool we possess to investigate reality, because it is based on continuous critical comparison.”

Reflections on scientific communication highlighted both its political and epistemic dimensions. Galileo’s vernacular publications were cited as politically motivated, while the correspondence between Mersenne and other scientists was seen as instrumental to knowledge construction. Students concluded that science should be understood not as a solitary endeavor but as a collaborative, communicative process.

From a pedagogical standpoint, students expressed appreciation for the cultural and philosophical framing of science. Many called for greater integration of NOS into standard curricula, not through isolated lectures but embedded in practical activities. A recurring theme was the civic importance of scientific literacy. One student emphasized, “It is important that the student, in addition to what science teaches, knows how science works, so that they can become a conscious and responsible citizen.”

In conclusion, students demonstrated a maturing epistemological understanding, recognizing science as tentative, empirical, creative, and socially embedded. Persistent ambiguities—especially concerning models, uncertainty, and the theory-laden nature of observation—underscore the need for instructional approaches that systematically integrate the history, philosophy, and practice of science to support a coherent conception of scientific knowledge.

7. Limitations of the Study

We acknowledge the limitations of our research that are common to qualitative investigations. First, as a case study focusing on a specific cohort of student teachers and pre-service teachers, the findings may not be directly generalizable to broader student populations (Lising & Elby, 2005), especially considering the specific type of students involved with a specific interest in educational aspects. Additionally, despite the use of multiple data sources and the comparison of different analysts’ interpretations (inter-rater reliability), strengthening the validity of our claims, the interpretation of semi-structured reflections remains subject to the researchers’ theoretical framing (Scherr & Hammer, 2009). Finally, the duration of the intervention may only capture the initial shifts in students’ complex epistemological structures, suggesting the need for longitudinal follow-ups.

8. Conclusions

The pedagogical activities described in the text have proven to be potentially effective both in covering the typical objectives of a physics laboratory course and in introducing and enhancing the understanding of many topics related to the NOS and the NOSI. The approach used, based on practical experiments inspired by a historical experiment of Galileo and integrated with historical reflections and explicit and reflective discussion on NOS/NOSI themes, aims to develop student teachers’ deeper understanding of how scientific knowledge is constructed.

Through experimental activity, students were able to confront the empirical nature of scientific knowledge, based on observation and experience, although the relationship between experiments and theoretical development may still present nuances to explore. They directly experienced the role of models as mediators between theory and data, understanding the processes of idealization and de-idealization and the predictive power of models. The activity also highlighted the inevitable uncertainty in science, distinguishing between aleatoric and epistemic uncertainty, and how this is not a limitation but an intrinsic characteristic.

Historical reflections, particularly on the case of Galileo and the figure of Mersenne, allowed for the deconstruction of the myth of the single scientific method, highlighting the plurality of approaches and the importance of creativity and subjectivity (such as theory-ladenness) in the scientific process. The crucial importance of the scientific community, communication and peer review as drivers of the self-correcting NOS emerged.

Although some conceptual challenges persist, such as the precise distinction between theories, laws, and models, the activity contributed to overcoming a naive view of science as a static set of absolute truths, promoting instead an understanding of its provisional, iterative, progressive and cumulative nature, fundamental for developing critical thinking and informed TIS.