Systematic morphological analysis and stylistic cultural interpretation of Song Dynasty Longquan Porcelain Vases

Abstract

The Longquan Porcelain Vases from China’s Song dynasty represent significant cultural heritage, yet their morphological characteristics and combinatorial patterns have lacked systematic investigation. This study integrates grounded theory with quantitative statistical analysis, examining 125 vase samples from 12 authoritative monographs. Open coding identified features, axial coding established hierarchical classifications, and Pearson correlation analysis revealed combinatorial patterns, while historical literature analysis ensured cultural validity. The study identified 7 major components, 22 morphological codes, and 68 significantly correlated feature combinations. Findings reveal that Longquan porcelain vase morphology exhibits “diversity within standardization,” manifested in systematic form combinations, unity of function and aesthetics, and associations with social hierarchy and cultural trends. These discoveries reveal Song dynasty Longquan porcelain craft standards and demonstrate their value as material culture carriers. This combined analytical approach provides empirical foundation for traditional craft research and new pathways for cultural heritage preservation.

Introduction

Porcelain art is an integral part of Chinese traditional culture, among which Longquan porcelain holds a distinctive position in Chinese porcelain history due to its unique forms and glazes¹. The Longquan Kiln, which began production during the Southern Dynasties, flourished in the Song Dynasty and declined in the Qing Dynasty, represents one of China’s longest-running and most influential porcelain production systems². During its developmental history, Longquan Kiln reached its artistic and technological pinnacle in the Song Dynasty. The works from this period featured elegant forms, exquisite craftsmanship, and fresh glaze colors, significantly advancing Chinese porcelain development and serving as crucial references for later porcelain production³. In September 2009, the traditional craftsmanship of Longquan porcelain was officially inscribed on the “ Representative List of the Oral and Intangible Heritage of Humanity,” becoming the world’s only porcelain item to receive this recognition⁴. Song Dynasty Longquan porcelains not only reflect the esthetic preferences of their time but also carry rich historical and cultural information. Among these, Porcelain Vases, as one of the most representative visual languages of the Longquan Kiln, demonstrate not only rich artistic variations but also clearly interpret the craftsmanship, esthetic preferences, and cultural connotations of Song Dynasty artisans.

Traditional porcelain art research has primarily relied on visual interpretation and personal experience. While this approach can capture the artistic qualities of porcelain, it has limitations in terms of systematicity and objectivity. In recent years, with the rise of interdisciplinary research methods, scholars have begun to apply more scientific and systematic approaches to porcelain art research. At the forefront of international research, Wang et al. proposed a systematic analytical framework based on elliptical Fourier descriptors, establishing a quantitative classification method for pottery morphological features through three-dimensional scanning technology, providing a more objective analytical pathway for traditional artifact research⁵. Sipiran et al. developed a data-driven analytical system that transcended the pottery morphological limitations of traditional visual observation through structured processing of point cloud data, achieving more precise and systematic feature extraction⁶. Rieke-Zapp and Trinkl established a standardized morphological measurement system, successfully implementing systematic comparative studies of ancient Greek porcelains, providing a repeatable and verifiable scientific analytical paradigm for porcelain art research⁷. These studies demonstrate the enormous potential of applying systematic quantitative analysis methods to traditional porcelain research.

The development of quantitative analysis techniques has opened new pathways for Chinese traditional cultural heritage research, including Chinese porcelain art. A series of systematic studies focusing on indigenous cultural heritage has demonstrated the unique value of this approach in revealing the inherent meanings of material culture. In the field of morphological quantitative analysis, scholars have conducted multidimensional quantitative research on traditional crafts by introducing typological methods. Research on the morphology of export cups from the Ming and Qing dynasties revealed an evolutionary trajectory in their design, transitioning from figurative representation in the Ming dynasty to abstract and minimalist styles in the Qing dynasty⁸, studies of export vessels from the Lingnan region, through precise measurements of vessel volume, proportions, and profile curves, demonstrated the intrinsic connections between morphological changes and functional use as well as social status⁹; analysis of Qing dynasty export porcelain bowls from Guangzhou further indicated that the widespread application of ring foot designs and the relative concentration of vessel dimensions reflected the systematic production modes and standardization trends of that time¹⁰. In terms of cultural visual representation and technical analysis, researchers have combined quantitative methods with artistic interpretation to explore new dimensions in cultural heritage research. Through iconographic methods analyzing facial expressions and costume colors in Qing dynasty export paintings, scholars have demonstrated specific manifestations of East-West esthetic fusion¹¹; color studies based on HSV color models and clustering algorithms revealed the transition from traditional Chinese formal color systems to Western realistic expression techniques¹²; while data mining analysis of ancestral hall paintings in Guangzhou successfully established association models between decorative arts and traditional ritual culture¹³. Regarding historical and cultural background research, scholars have revealed the profound influence of regional cultural integration and national policy intervention on the formation and evolution of traditional crafts by examining the development of decorative industries in early port cities of the South China Sea region and the historical changes in porcelain production in Jingdezhen^14,15. These studies not only demonstrate the applicability of quantitative methods in traditional craft research but also provide important references for expanding the methodological system of cultural heritage research.

Within systematic quantitative analysis methods, grounded theory, as a structured qualitative research method, is particularly suitable for in-depth exploration of complex cultural phenomena. This method effectively captures the intrinsic connections between artifact features through systematic coding processes that distill concepts and construct theoretical frameworks from data and has gained increasing application in cultural heritage research in recent years¹⁶. Yan and Sun applied grounded theory to analyze the cultural ecological system of Chinese She ethnic minority costume, successfully constructing a systematic analytical framework for understanding cultural heritage values¹⁷. Bakri et al. used grounded theory to study the relationship between local communities and cultural heritage in George Town World Heritage Site, Malaysia, revealing deep connections between World Heritage Site designation and local communities¹⁸. At the methodological level, the application of grounded theory has been continuously developing and transforming. Seaman explored the possibility of flexibly integrating grounded theory as a research method with existing theoretical frameworks, suggesting that it can transition from a methodology with positivist foundations to a research method usable within different theoretical frameworks, providing new analytical perspectives for cultural historical research¹⁹. Barnes emphasized the special challenges of applying grounded theory in cross-cultural research, pointing out that researchers must pay particular attention to the perspectives of the studied culture and establish sufficiently rich inductive theories through constant comparison of emerging concepts from cultural perspectives²⁰. Meanwhile, the integration of grounded theory with other research methods has become an important development trend in cultural heritage research. Charmaz and Belgrave critically examined the status of data in grounded theory research against the backdrop of neoliberalism, emphasizing the importance of methodological self-awareness, which provides insights for balancing qualitative and quantitative methods in cultural heritage research²¹. Guetterman et al.’s systematic analysis of the integration of mixed methods and grounded theory revealed the widespread application of this methodological integration across different disciplines, providing procedural guidance for method integration, although they also pointed out the current lack of theoretical development in research²². Although applications of grounded theory in porcelain art research are still rare, its successful applications in cultural heritage research provide valuable reference.

Though existing research has made significant progress in the application of quantitative methods, systematic review of relevant literature reveals several directions worth exploring in the study of the Song dynasty Longquan kiln porcelain morphological system. First, existing research often employs single-dimensional quantitative measures (such as geometric morphometrics or color analysis), leaving room for development in integrating multi-dimensional feature analysis. While these methods have made breakthroughs in specific dimensions, the description of complex vessel overall morphological systems can be further improved. Second, the integration of quantitative research and cultural interpretation is an ongoing exploratory process. Technology-oriented morphological analysis provides precise data description, while cultural interpretation offers in-depth understanding, with the dialog mechanism between the two still continuously developing. Additionally, in terms of data acquisition and processing, existing research largely relies on high-precision data collected by modern instruments, while methods for processing information from historical literature and image materials also need continuous innovation, especially for studying traditional porcelain with extensive history such as Song dynasty Longquan kiln products. From a research perspective, besides focusing on the connection between vessel morphology and external social factors, further exploration of the organizational logic and cultural connotations within the vessel’s internal morphological system is also an important approach to expanding understanding of traditional craft systems. The development of these research directions will help gain a more comprehensive grasp of the complexity and multi-dimensionality of traditional porcelain art.

Based on the recognition of limitations in existing research, this study selects Song dynasty Longquan kiln porcelain vases, which possess rich morphological features and typical cultural value, as the research subject, proposing a research framework combining grounded theory with quantitative statistical analysis. The research selects porcelain vase samples from authoritative monographs, conducts systematic coding and correlation statistical analysis according to this framework, and explores the internal patterns and cultural connotations of the Longquan porcelain vase morphological system by integrating historical literature interpretation. This combination of qualitative and quantitative methods demonstrates clear scientific validity and research relevance. From a methodological scientific perspective, grounded theory achieves systematic construction from raw data to concepts to theory through the progressive process of open coding, axial coding, and selective coding, ensuring the objectivity of morphological feature extraction and the rigor of the classification system; while Pearson correlation analysis, based on statistical principles, provides quantitative verification of morphological feature combination patterns under conditions controlling for other variables, making the research results replicable and verifiable. From a research relevance perspective, this methodological integration effectively bridges the methodological gap between qualitative description and quantitative analysis in traditional cultural heritage research, capturing subtle differences and cultural symbols in vessel morphology through systematic coding, while revealing inherent association patterns between morphological elements through statistical tools. This approach is particularly suitable for studying traditional crafts with complex morphological systems and profound cultural connotations, providing a more comprehensive analytical framework for understanding cultural heritage that combines technical sophistication and artistic value, such as Song dynasty Longquan kiln porcelain vases. The main tasks of this study include: (1) establishing a multi-dimensional feature classification system, systematically coding and classifying the morphological features of Song dynasty Longquan kiln porcelain vases through grounded theory; (2) revealing the distribution patterns of morphological features, analyzing the distribution proportions and types of morphological features of each component; (3) exploring morphological combination patterns, determining the combination relationships between morphological features using Pearson correlation analysis; (4) interpreting cultural and social significance, analyzing the cultural background and social functions of morphological features in combination with historical literature. This study aims to deepen the understanding of Song dynasty Longquan kiln porcelain craftsmanship while providing a methodological approach that can be referenced for systematic research on other traditional crafts.

Methods

Materials

This study selected appropriate hardware equipment and software tools based on research requirements. In terms of hardware, a Dell Precision 7920 Tower workstation equipped with an Intel Xeon Gold 6258R processor and 128 GB RAM was used for data processing. A Canon EOS 5D Mark IV camera was employed for high-resolution image acquisition. For software applications, NVivo 14 Pro from QSR International was utilized for grounded theory coding work. Research by Hutchison et al. has confirmed this software’s effectiveness in facilitating systematic grounded theory research²³. Additionally, IBM’s SPSS Statistics 26 was used for statistical analyses including Pearson correlation analysis. Research by Xi et al. has demonstrated this software’s significant value in multivariate statistical analysis of cultural heritage²⁴. Microsoft Excel 2019 was also used for organizing and preliminary analysis of raw data.

The sample data for this study primarily originated from published monographs and catalogs on Song Dynasty Longquan Porcelain Vases. Initially, 32 relevant monographs were collected through various channels, including libraries, online e-book stores, and physical bookstores. In the sample selection process, several key elements were evaluated. (1) Content relevance: The monographs had to focus primarily on Song dynasty Longquan kiln porcelain, especially vase-type vessels, and provide a sufficient number and variety of examples of Song dynasty Longquan porcelain vases. (2) Comprehensiveness of textual and visual descriptions: The selected monographs needed to include high-quality images of the porcelain, as well as detailed textual descriptions covering key information such as morphological features, dimensions, glaze colors, and decorative patterns of the porcelain vases. (3) Clarity of archeological background: Priority was given to monographs that provided clear excavation information, collection sources, or authentication certifications to ensure the authenticity and accuracy of the research subjects. (4) Professional authority of the authors: The academic backgrounds, research experiences, and reputations of the monograph authors in the field of Longquan kiln porcelain research were evaluated, with priority given to monographs written by renowned porcelain research experts, archeologists, museum scholars, or experienced collectors to ensure the academic reliability of the content. After rigorous screening, 12 monographs were ultimately selected as research samples, including “Longquan Porcelain,” “Ancient Chinese Porcelains—Longquan Kiln,” “Art of Longquan Porcelain: Jade-like Brilliance,” “Identification of Longquan Kiln,” “Appreciation of Imperial Longquan Porcelain,” “Identification and Appreciation of Longquan Kiln,” “Research on Longquan Porcelains,” “Comprehensive Mirror for Longquan Kiln,” “Early Green of Plums: Longquan Porcelain,” “Ancient Chinese Famous Kiln: Longquan,” “Chinese Famous Kiln Sites Series: Longquan Kiln,” and “Beloved Creations of Kua Cang: Selected Longquan Porcelain.” From these monographs, approximately 200 potential Song dynasty Longquan porcelain vase cases were initially identified, then based on criteria such as completeness of information, clarity of morphological features, and era typicality, 125 representative cases were ultimately selected for in-depth study. The distribution of these samples across the various monographs was not uniform, with certain authoritative texts providing more research cases due to their comprehensiveness and representativeness, while other monographs served as supplementary and verification materials.

This literature-based sampling method holds significant value in porcelain art research. Wu and Zou emphasized the central role of literature-based data selection in porcelain archeology and art research²⁵. To further ensure the objectivity and rigor of the collected data, multiple measures were taken to address potential issues. (1) Cross-validation: Occurrences of the same porcelain in different literature sources were cross-compared to obtain more comprehensive information descriptions. (2) Expert consultation: When information was unclear or ambiguous, porcelain research experts were consulted to obtain professional judgments and supplementary information. (3) Information integration: Since some original texts were not sufficiently complete, to ensure analytical accuracy, important information expressed in the accompanying images was manually supplemented. Specifically, two experts separately confirmed the image information and, after reaching consensus on the missing content, transformed it into textual information to supplement the text. (4) Multi-source corroboration: During the coding process, judgment was based not only on the images themselves but also on relevant image materials and literature to corroborate and ensure consistency between images and textual descriptions. Ultimately, 125 Song dynasty Longquan kiln porcelain vase samples were successfully obtained. Each case contains multidimensional information including porcelain vase photographs, morphological descriptions, and glaze characteristics. Zhang and Zheng noted that multidimensional data collection helps form a more comprehensive understanding of cultural heritage²⁶. Pavlidis et al. also emphasized that data recording and analysis of cultural heritage requires consideration of multiple dimensions²⁷. Therefore, particular attention was paid to the data distribution across different dimensions in the samples to ensure the research could comprehensively reflect the overall characteristics of Song Dynasty Longquan Porcelain Vases.

Grounded theory coding

Grounded theory, first proposed by Glaser and Strauss in 1968, is a qualitative research method that systematically generates theory from data²⁸. This method is particularly suitable for exploratory research, helping researchers distill key concepts and patterns from large amounts of data. In this study, the grounded theory coding method developed by Strauss and Corbin in 2014 was employed, which primarily comprises three stages: open coding, axial coding, and selective coding²⁹.

In the open coding stage, the descriptions of Longquan Porcelain Vases from selected monographs were analyzed, with initial codes assigned to identified feature concepts. For example, when describing vase mouth shapes, concepts were extracted such as “straight mouth,” “flaring mouth,” and “contracted mouth.” The axial coding stage focused on analyzing logical relationships between concepts, categorizing related concepts into higher-level categories. Similarly, for the neck component, we identified multiple sub-types such as “short cylindrical neck,” “long cylindrical neck,” and “long constricted neck.” These distinctions were based not only on textual descriptions from authoritative monographs but also on visual morphological inspection of high-resolution images. Specifically, “long cylindrical neck” is defined by a vertical neck section exceeding one-third of the vessel’s height with relatively parallel sides, while “long constricted neck” maintains a similar length but features a pronounced inward curvature resembling a bottleneck. This differentiation reflects both geometric proportion and contour variation. When descriptive texts were ambiguous, two researchers independently cross-validated the morphological feature using image evidence to ensure consistency. This classification approach was designed to maintain theoretical sensitivity while ensuring conceptual clarity and reproducibility in feature coding. For instance, concepts like “round lip” and “square lip” were categorized under the higher-level category of “lip morphology.” In the selective coding stage, attempts were made to identify core categories and relate all other categories to them. Throughout the coding process, particular attention was paid to maintaining theoretical sensitivity, a key element of grounded theory research emphasized by Glaser³⁰. Different concepts and categories were continuously compared, with thought processes documented through memos to ensure the analysis closely aligned with the data while elevating to a theoretical level.

Saturation testing

In grounded theory research, theoretical saturation is a key concept, referring to the state reached when new data no longer provides additional insights for theory development. Saunders et al. emphasized the importance of evaluating saturation across multiple levels, including coding, themes, and theory³¹. To ensure the coding process achieved theoretical saturation, a systematic saturation testing method was employed. First, detailed grounded theory coding was conducted on 60% of the selected cases of Song Dynasty Longquan porcelain vases to establish an initial coding framework. Second, the remaining 40% was sequentially coded, with an evaluation after each case to determine whether new concepts or categories emerged. If all newly emerging concepts could be classified within existing categories, this indicated that the coding framework had reached saturation. To further ensure coding reliability, inter-coder reliability testing was employed³². 30% of the sample data was selected for double coding, a proportion higher than the general level in this field of research. Two researchers with backgrounds in porcelain art research were invited to independently complete the coding work. The consistency between coders was calculated, comparing whether it exceeded the acceptable standard of 70%. For coding discrepancies, discussions were held to clarify the causes of disagreement and refine the coding standards accordingly.

Pearson correlation analysis

After completing grounded theory coding and saturation testing, Pearson correlation analysis was employed to explore relationships between different morphological features of Song Dynasty Longquan Porcelain Vases. Pearson correlation analysis is a statistical technique used to measure the relationship between two variables while controlling for the effects of other variables³³. This statistical method has demonstrated significant value in cultural heritage research. For example, Chen et al. successfully identified intrinsic connections between visual environmental features when studying the combination of landscape elements in Chinese cultural heritage gardens using Pearson correlation analysis³⁴. In practical implementation, the coded morphological features were first converted into quantifiable variables. For instance, features such as “lip morphology” were encoded as binary variables (0 indicating absence, 1 indicating presence). Using IBM SPSS Statistics 26 software, the analysis process included data preprocessing, variable selection, significance testing, and result interpretation. In interpreting the results, consideration was given not only to the significance between variables but also to the actual magnitude of correlation coefficients, as well as the different morphological patterns potentially reflected by positive and negative correlations.

To further assess the structural characteristics of the dataset and evaluate potential multicollinearity among morphological features, a principal component analysis was conducted using the binary-coded variables. Principal component analysis served as a dimensionality-reduction technique to identify major variance-bearing components and detect feature redundancy within the morphological system. Principal component analysis inherently identifies multicollinearity by transforming correlated original features into a new set of uncorrelated principal components. Features contributing minimal variance to principal components indicate redundancy or collinearity. By identifying the most significant principal components, we determined which morphological features carried unique variance and which exhibited high intercorrelations^35,36.

Through Pearson correlation analysis, multiple sets of significantly correlated morphological feature combinations were identified, such as the correlation between dish-shaped mouth and fish-shaped handles. These quantitative results not only verified findings from qualitative analysis but also provided more precise relationship descriptions. However, it was also recognized that correlation does not imply causation; therefore, in interpreting these results, historical literature and expert knowledge were integrated to ensure the validity of interpretations at both statistical and historical-cultural levels. This combination of quantitative analysis with qualitative analysis from grounded theory helped achieve a more comprehensive and objective understanding of the morphological patterns of Song Dynasty Longquan Porcelain Vases.

Method integration and research framework

The integration of grounded theory and statistical analysis in this study reflects a methodological shift in the research paradigm of traditional crafts. This integration is not merely a sequential use of qualitative and quantitative methods, but rather constructs a cyclically interactive cognitive system, with its theoretical foundation derived from Creswell’s “explanatory sequential design” and Morgan’s “priority-implementation model”^37,38. In this study, the priority of grounded theory is manifested in its inductive logic for the conceptualization of porcelain morphological features, providing a categorical basis for subsequent statistical analysis; while Pearson correlation analysis extends the theoretical construction capability of grounded theory, objectively testing morphological association patterns through verification logic. This methodological integration breaks down the barriers between qualitative description and quantitative analysis in traditional porcelain research, achieving a spiral deepening process of “qualitative discovery-quantitative verification-qualitative interpretation.” Compared with single methods, this integration offers three key advantages: first, compared to purely qualitative research, it increases the objective verification of morphological feature associations; second, compared to purely statistical analysis, it enhances the cultural sensitivity of morphological variable extraction; most importantly, this integration establishes a reasoning pathway from material features to cultural connotations, enabling morphological analysis and cultural interpretation to form mutual verification on an empirical basis. This methodological integration strategy has special applicability for traditional craft research that combines technical characteristics with cultural connotations, providing a new methodological perspective for traditional porcelain research.

Based on the aforementioned methodological integration approach, a systematic research framework was constructed, as shown in Fig. 1. This flowchart displays the complete research process from sample collection to final results, clearly presenting the specific steps at each research stage. In the sample collection phase, 12 monographs were selected from 36 relevant works based on criteria such as content relevance, completeness of textual and visual descriptions, and clarity of archeological background, ultimately extracting 125 Song dynasty Longquan kiln porcelain vase cases. The feature coding phase followed a progressive process of open coding, axial coding, and selective coding, with consistency testing to ensure coding reliability. The quantitative analysis phase included steps such as code conversion, data preprocessing, variable selection, significance testing, and result interpretation, identifying inherent associations between morphological features by controlling for variable influences. This framework ultimately helped identify 7 main components, 22 original morphological codes, and 68 significantly correlated feature combinations, providing a scientific foundation for exploring the internal patterns of the Longquan kiln porcelain vase morphological system.

Fig. 1

Flowchart of the research process.

Results

Multi-dimensional feature framework of Longquan Porcelain Vases

Through systematic grounded theory coding analysis, a comprehensive multi-dimensional feature framework for Longquan Porcelain Vases was constructed. During the open coding stage, detailed coding of the Song Dynasty Longquan porcelain vase case data from selected monographs was conducted, identifying a total of 1052 original concepts. Through repeated comparison and refinement, 88 open codes were derived, with examples of the open coding process shown in Table 1.

Table 1 Examples of open coding process.

In the axial coding stage, the relationships among these 88 initial concepts were further analyzed, organizing them into higher-level categories. Through this process, 20 subcategories and 4 main categories were derived. These categories formed the primary structure of the feature framework, reflecting the main aspects of Song Dynasty Longquan porcelain vase morphological characteristics. The process of axial coding and partial examples are shown in Table 2, which presents how open coding was systematically categorized into corresponding subcategories and ultimately formed into main categories. To more intuitively display the coding data flow and category formation process, Fig. 2 provides a flowchart from initial coding to subcategory identification, clearly presenting how different codes converge to form a higher-level classification system, as well as the proportional distribution of each category in the overall sample.

Fig. 2

Flow of axial coding from initial coded data to identified subcategories.

Table 2 Examples and results of axial coding process

In the selective coding stage, “Multi-dimensional Features of Song Dynasty Chinese Longquan Porcelain Vases” was identified as the core category. This core category integrated all main categories, reflecting the central theme of the research. After completing all coding work, systematic saturation testing methods were employed to ensure the coding process had reached theoretical saturation. Through inter-coder reliability testing, it was found that consistency reached 87%, exceeding the acceptable standard of 70%, and the coding standards were refined based on the causes of disagreements. Based on the final coding results, a multi-dimensional feature framework for Longquan Porcelain Vases was constructed, as shown in Fig. 3. From the overall coding results and multi-dimensional feature framework, descriptions of Song Dynasty Chinese Longquan Porcelain Vases primarily focused on morphology, glaze color, and decorative patterns. Among these, morphological descriptions occupied the largest proportion, which aligns with the historical assessment that Longquan porcelains excel in form. Considering the central position of morphological features in Longquan Porcelain Vases and the focus of this research, subsequent analysis concentrated on studying the distribution proportions and combinatorial patterns of morphological features.

Fig. 3

Multi-dimensional feature framework of Song Dynasty Chinese Longquan Porcelain Vases.

Distribution system of morphological feature types in Longquan Porcelain Vases

The morphological features of Longquan Porcelain Vases demonstrate rich diversity, reflecting the high level of porcelain craftsmanship development during the Song Dynasty. Through systematic analysis, the distribution of feature types for each component was derived, as shown in Table 3.

Table 3 Morphological distribution of Longquan Porcelain Vases

Lip morphology primarily comprises two types: square lip and round lip. Square lips feature right-angled or near-right-angled straight edges, while round lips have curved or circular-shaped edges. Among all samples, square lips account for 58.82% and round lips for 41.18%. The round lip structure is better suited for direct contact with the human mouth, thus appearing more frequently on small vases and vessels such as bowls and plates intended for drinking, reflecting the practical design tendencies of Song Dynasty porcelains. Although square lips are less practical for drinking purposes than round lips, they present a more dignified and formal esthetic appearance and demand higher craftsmanship skills, better demonstrating the exquisite artistry of the Longquan Kiln. Consequently, square lips were more widely applied to porcelain vases not intended for drinking purposes. These distribution characteristics provide important clues for understanding the technical capabilities and esthetic preferences of the Song Dynasty Longquan Kiln.

Mouth morphology exhibits more diverse variations, with eight main types identified: straight mouth, dish-shaped mouth, bowl-shaped mouth, flaring mouth, contracted mouth, trumpet mouth, foliated mouth, and outward-slanting mouth. The straight mouth features a vertical cylindrical shape. The dish-shaped mouth resembles a plate with straight walls and inward folding. The bowl-shaped mouth has a slightly outward-flaring rim with a relatively small diameter compared to the vase body, slightly thickened lip, and smooth transitional curves. The flaring mouth gradually widens toward the rim. The contracted mouth gradually converges inward at the rim. The trumpet mouth gradually expands from a thin neck, resembling a wind instrument’s bell. The foliated mouth divides the circular or oval opening into several equal segments with continuous arcs, resembling flower petals. The outward-slanting mouth features an outward-turned rim with a slight trumpet-like appearance. Among mouth types, straight mouths are most common, accounting for 45.8%, followed by dish-shaped mouths at 25.4%. Flaring, contracted, and bowl-shaped mouths each represent over 5% of the sample. The straight mouth design became predominant due to its simplicity and wide applicability; this vertical cylindrical opening facilitated sealing and storage of the porcelain vases while aligning with the Song Dynasty’s esthetic preference for understated elegance. The dish-shaped mouth, with its outward-flaring rim and flowing curves, creates an elegant vessel profile that was highly valued in contemporary display settings³⁹. Although less common than straight and dish-shaped mouths, other types such as flaring, contracted, bowl-shaped, trumpet, and foliated mouths fulfilled diverse functional requirements while demonstrating unique esthetic values. Particularly, foliated and trumpet mouths, with their exquisite and distinctive designs, were primarily used in high-end premium porcelains, symbolizing Longquan artisans’ artistic pursuit and innovation.

Neck morphology refers to the structural design and shape characteristics between the vase mouth and shoulder, serving as a transitional and connecting component in the overall vessel form. In the statistical results, neck morphology primarily includes short cylindrical neck (31.6%), long cylindrical neck (27.3%), and long constricted neck (24.2%). The prunus vase is a representative example of the short cylindrical neck, featuring a small mouth, broad shoulders, thin base, and ring foot, as shown in Fig. 4a⁴⁰. It is named as the rim is so small that only plum twigs could be inserted in, and it is well-known as a container for liquor⁴¹. The long cylindrical neck was imbued with symbolic meaning of “longevity” by Song Dynasty scholars⁴². The pure vase is a typical representative of the long cylindrical neck design in Song Dynasty Longquan Porcelain Vases⁴³. Originally used by wandering monks for drinking water and hand washing, water droppers gradually entered the daily life of Song Dynasty people. Many water droppers feature a protruding rim on their long cylindrical necks, making the overall vessel shape resemble the Chinese character “ji” (吉, meaning auspicious), hence also known as the “great fortune vase,” symbolizing great prosperity, as shown in Fig. 4b. Another representative of the long cylindrical neck is the paper-mallet vase, named after its resemblance to the paper-making tool “paper mallet,” with its long cylindrical neck mimicking the mallet’s handle, as shown in Fig. 4c. The paper-mallet vase features well-proportioned long neck and full body, creating a vertical and upright visual effect, making it a favored decorative piece among Song Dynasty scholars.

Fig. 4: Representative examples showing neck morphology of Song Dynasty Longquan Porcelain Vases.

a Carved celadon meiping vase, Longquan ware, Southern Song Dynasty (1127–1279), collection of Longquan Celadon Museum. Source: Siyuwj, Wikimedia Commons, licensed under CC BY-SA 4.0. b Celadon vase with “Great Luck” inscription, Longquan ware, Southern Song Dynasty (1127–1279). Source: Siyuwj, Wikimedia Commons, licensed under CC BY-SA 4.0. c Celadon vase with dish-shaped mouth, Longquan ware, Zhejiang province, Southern Song Dynasty (1127–1279 A.D.). Source: Rosemania, Wikimedia Commons, licensed under CC BY 2.0.

Shoulder morphology primarily includes three types: sloping shoulder, angular shoulder, and full shoulder. Among these, sloping and angular shoulders account for 32.8%, while full shoulders represent 31.3%. The sloping shoulder design features a downward slope that transitions into the belly in a curved arc. The angular shoulder design shows a distinct transition from shoulder to belly, creating an external angle at the transition point. In contrast to the angular shoulder, the full shoulder design features rounded, full contours that rise upward, making the overall porcelain vases more voluminous while also increasing the capacity of the vase body.

The belly, being the main storage space in the middle section of porcelain vases, represents the most diverse part in the overall vase form, demonstrating the variety of Longquan porcelains. Belly morphology primarily comprises three main types: curved belly, cylindrical belly, and flattened belly. Among these, curved belly designs account for 39.2%, featuring an outward protrusion with full, swelling contours, making it the most common belly type. Cylindrical belly designs represent 22.5%, featuring straight or near-straight cylindrical forms that appear clean and neat, commonly used in more slender porcelain vases. Flattened belly designs account for 8.82%, with an oblate cross-section where the front-to-back distance is less than the side-to-side width. Although flattened bellies are less common than curved and cylindrical types, they present significant technical challenges during firing, thus reflecting the sophisticated craftsmanship of the Longquan Kiln, and have become one of the classic designs of Song Dynasty Longquan Porcelain Vases. Other belly types include spherical belly, gallbladder-shaped belly, lobed belly, egg-shaped belly, square columnar belly, and olive-shaped belly. Notably, during the Song Dynasty, there was a prevalent trend of imitating antiquities, with jade-inspired vessels like the cong-shaped vase being particularly popular, as shown in Fig. 5a. Consequently, the square columnar belly of cong-shaped vases became a distinctive belly design popular during the Song Dynasty.

Fig. 5: Representative examples of Song Dynasty Longquan Porcelain Vases.

a Vase in the shape of a Neolithic jade cong, China. Source: Unknown author, Wikimedia Commons, dedicated to the public domain (CC0). b Bottle with lid, Longquan ware, Northern Song Dynasty (960–1127), Dallas Museum of Art. Source: Daderot, Wikimedia Commons, dedicated to the public domain (CC0).

Statistical analysis of foot morphology shows that hollow ring feet hold an absolute dominant position, accounting for 98.4% of all cases. This design was widely adopted for three main reasons. First, hollow ring feet facilitate more even heat distribution during firing, reducing the risk of cracking or deformation, thus improving the success rate of production. Second, the hollow design reduces the overall weight of the porcelain vases, making the vase lighter and more convenient for daily use and handling. Additionally, hollow ring feet conserved raw materials and improved production efficiency, meeting the mass production requirements of the Longquan Kiln during that period.

In handle morphology, the main types include handles-free, two loop handles, pierced handles, phoenix-shaped handles, dragon-shaped handles, and fish-shaped handles. In the research samples, 85.6% of Longquan Porcelain Vases feature no handles, reflecting the Song Dynasty esthetic preference for simplicity and understated elegance. Among vases with handles, two loop handles account for 44.4%, making it the most common handle type. The widespread use of two loop handles is due to their functionality in facilitating transportation and hanging while not disrupting the vessel’s overall contours. For example, prunus vases used for storing liquor often feature two loop handles on the shoulders to facilitate carrying, as shown in Fig. 5b. Although dragon-shaped, phoenix-shaped, and fish-shaped handles with their more decorative nature appear in lower proportions, their presence enhances the vases’ ornamental value⁴⁴. The phoenix-shaped handle design first appeared on Song Dynasty Longquan Porcelain Vases, while fish-shaped handles were found on both Song Dynasty Longquan and Ge Kiln vases. These designs continued into the Qing Dynasty, as seen in examples such as the Kangxi purple-glazed phoenix-handled garlic-mouth vase, the Qianlong porcelain blue-and-white phoenix-handled vase, and the Kangxi wucai fish-handled vase with auspicious sea creatures pattern. This not only demonstrates the Song Dynasty Longquan Kiln’s pursuit of design innovation but also had a profound influence on later porcelain morphology, establishing a unique esthetic legacy in porcelain art.

Through systematic coding, this study identified 22 morphological elements across 7 major components of Song Dynasty Longquan Porcelain Vases and constructed a morphological feature distribution system, as shown in Fig. 6. The figure employs a concentric circle hierarchical structure to display the classification system from inner to outer layers: the innermost circle represents the core concept of “Song Dynasty Longquan Porcelain Vase Morphology,” extending outward to encompass seven major components and their subordinate specific types. Different colors distinguish each major component, while the size of dots reflects the frequency of occurrence for each type in the sample. This distribution system provides a fundamental framework for subsequent research into the combinatorial patterns of morphological features across different components.

Fig. 6

Morphological feature distribution system of Song Dynasty Longquan Porcelain Vases.

Morphological comparison between Northern and Southern Song Dynasties

Based on the morphological feature coding, this study conducted a period analysis of Longquan kiln porcelain vase samples, aiming to reveal the diachronic changes of morphological features within the Song dynasty and address the research need for associations between porcelain morphological evolution and historical background. Among the 125 samples, according to chronological information provided in the literature, 120 samples with more detailed time annotations were identified, including 60 samples from the Northern Song dynasty (960–1127) and 60 samples from the Southern Song dynasty (1127–1279), with 5 samples excluded from the analysis due to lack of clear period classification information. To scientifically present the distribution differences of morphological features between the Northern and Southern Song periods, frequency statistics were performed on the morphological coding data for both periods, calculating the occurrence proportion of each morphological type within their respective period samples. Figure 7 displays the statistical results in the form of an area chart, where the horizontal axis represents different morphological feature types, the vertical axis represents the occurrence proportion, the red area represents the feature distribution of Northern Song samples, and the blue area represents the feature distribution of Southern Song samples.

Fig. 7

Comparison of the distribution of morphological features of various parts of Longquan kiln porcelain bottles from the Northern and Southern Song dynasties.

Analysis of Fig. 7 reveals that the Longquan kiln porcelain vase morphological system exhibits both commonalities and differences between the Northern and Southern Song periods. Regarding lip morphology, square lips maintained a high proportion in both periods, demonstrating continuity in porcelain-making traditions; however, mouth morphology showed significant changes, with straight mouths being the dominant feature in the Northern Song, while the proportion of dish-shaped mouths increased significantly in the Southern Song. This change may be related to the expansion of vessel functions, with dish-shaped designs enhancing use diversity while maintaining esthetic value. Neck morphology also presented distinct period characteristics, with short cylindrical necks dominating the Northern Song period, creating minimalist forms; while the application of long cylindrical necks notably increased in the Southern Song period, giving vessels a more elongated visual effect. Shoulder and belly designs similarly reflected period differences, with the Northern Song primarily featuring full shoulder and curved belly combinations, forming stable vessel shapes; while the Southern Song saw an increased proportion of sloping shoulders, which, paired with curved bellies, created more fluid lines. Regarding handle design, both periods primarily featured handleless designs, but the proportion of decorative handles (such as phoenix-shaped and dragon-shaped handles) increased in the Southern Song, reflecting enhanced decorative considerations. Notably, hollow ring feet maintained absolute dominance in both the Northern and Southern Song periods, reflecting the stable inheritance of Longquan kiln porcelain-making techniques. By analyzing the most dominant morphological types for each component—those that not only exhibit the highest proportional frequency but also meet a threshold of ≥20% occurrence—typical morphological combinations for both periods were identified. As shown in Fig. 8, the left side illustrates the typical Northern Song combination (square lip, straight mouth, handleless, short cylindrical neck, full shoulder, curved belly, hollow ring foot), while the right side depicts the typical Southern Song combination (square lip, dish-shaped mouth, handleless, long cylindrical neck, sloping shoulder, curved belly, hollow ring foot). These two sets of representative morphological combinations not only provide important chronological reference standards for dating Longquan kiln porcelain vases from the Song dynasty, but also visually demonstrate the diachronic evolution of Longquan porcelain-making techniques—from functionality toward esthetic refinement.

Fig. 8

Schematic Diagram of Typical Morphological Compositions of Longquan Kiln Ceramic Bottles from the Northern and Southern Song Dynasties.

The differences in the morphological characteristics of ceramic bottles from the Longquan kilns during the Northern and Southern Song dynasties are closely related to the unique social environments of those periods. From a political perspective, the Northern Song period was characterized by stable centralized authority and national unity, with official regulations guiding craft production. The various handicraft management regulations preserved in the “Collected Documents of Song Dynasty Institutions” reflect the strict regulatory environment of this period, which likely provided the institutional foundation for the standardized features observed in ceramic forms⁴⁵; After the Southern Song regime relocated to Lin’an, the Longquan kilns benefited from the new esthetic demands of the scholar-official class, gaining broader space for innovation⁴⁶, With the comprehensive prosperity of the Jiangnan economy during the Southern Song period, markets became increasingly segmented. The diverse urban life depicted in A Record of Dreamlike Life (Mengliang Lu) suggests increasingly complex consumer demands, which likely propelled ceramic forms toward greater stylistic diversity. The use of luxury materials such as gold and silver vessels across different temporal, spatial, and social contexts demonstrates a degree of regional variation in material culture practices. This provides valuable insight into the localization and functional transformation of ritual vessels, highlighting how vessel morphology in ceremonial settings simultaneously conveyed both practical and symbolic meanings of social identity^47,48. Cultural atmosphere transitions provided the deeper impetus for morphological evolution—The “ Treatise on Architectural Methods or State Building Standards” of the Northern Song period reflected a societal orientation toward order, rationality, and standardization in the technological and cultural environment⁴⁹, with designs emphasizing functionality and standardization, this pragmatic spirit directly reflected in the regular shapes and simple decorations of Longquan ceramics; By the Southern Song dynasty, with the increasing influence of the scholar-official group, the literary gatherings and daily life scenes recorded in “Old Stories of Wulin.” demonstrate how objects became symbolic carriers of cultural identity. This cultural shift was likely the intrinsic reason for Southern Song Longquan ceramic bottles’ greater emphasis on artistic expressiveness⁵⁰. Archeological evidence supports this viewpoint—comparative findings reveal that Longquan ceramics unearthed from Southern Song scholar-official tombs typically feature more elaborate and complex forms, while similar objects from ordinary civilian tombs are relatively simple, clearly revealing the corresponding relationship between social class and object morphology.

Through a systematic comparison of the morphological characteristics of Longquan kiln ceramic bottles from the Northern and Southern Song dynasties, a trend can be observed developing from function-dominated design toward a balance between esthetics and functionality: Northern Song period morphological designs emphasized practicality and production efficiency, displaying regularized and standardized characteristics; Southern Song period designs, while maintaining basic functionality, further strengthened artistic expressiveness and cultural connotations, exhibiting diversified and refined morphological features. This evolution reflects both the continuous advancement of ceramic craftsmanship and the profound transformation of social structure and cultural concepts during the Song dynasty. However, it should be noted that these morphological characteristics did not exist in isolation, but rather constituted complete vessel styles through specific combinatorial relationships. The next section will systematically analyze these characteristic combination patterns and their cultural implications.

Morphological combination patterns and cultural interpretation of Longquan Porcelain Vases

The morphological language of Song Dynasty Longquan Porcelain Vases exhibits systematic multi-dimensional characteristics, with vessel structures demonstrating both the inheritance of traditional porcelain craftsmanship and Song Dynasty artisans’ innovative exploration of proportional relationships⁵¹. Notably, the esthetic value of Song Dynasty Longquan Porcelain Vases is not merely manifested in the refinement of individual components, but rather through the dimensional coordination and rhythmic variations among different parts, creating a unique overall esthetic appeal⁵². To analyze the internal patterns of these formal combinations in depth, this study employed Pearson correlation analysis to conduct quantitative research on the collected sample data, aiming to reveal multiple statistically significant morphological combination patterns. Specifically, in this study, P < 0.05 indicates significant correlation, with correlation coefficients >0 indicating positive correlation and <0 indicating negative correlation. The following analysis is based on these criteria.

Prior to identifying significant combinations of morphological features, we applied principal component analysis to reveal the contribution of the features to the overall variance. The principal component analysis results showed that the first principal component accounted for 9.74% of the total variance, and the second component explained 7.59%. The first principal component is driven primarily by handle and foot morphology: handle-free (loading = +0.4055), two-loop handles (loading = –0.3894), solid ring feet (loading = –0.3578) and hollow ring feet (loading = +0.3578). Thus, the first principal component captures the contrast between handle-free vessels with hollow-ring feet and vases featuring two-loop handles and solid-ring feet. The second component separates vases with straight mouths (loading = –0.2790) from those combining long cylindrical necks (loading = +0.4278), dish-shaped mouths (loading = +0.2767) and hollow-ring feet (loading = +0.2533) (Fig. 9). Together, the top five components explained 36.11% of the variance, and the top ten accounted for 58.37% (Fig. 10). This distribution indicates that the morphological features are widely dispersed across multiple orthogonal dimensions, reflecting a high degree of structural diversity. From the 35th component onward, standard deviations declined sharply (e.g., 0.08%), indicating negligible variance and potential redundancy among a small number of binary morphological traits. These results confirm that while some collinearity exists—as expected in morphological systems—it is not overwhelming, and the dataset retains a high degree of informational integrity suitable for correlation-based analysis.

Fig. 9

Contribution of vase morphological features in the principal component space.

Fig. 10

Variance explained by each principal component derived from binary-coded morphological features.

The lip and mouth are closely connected, and as important components of the porcelain vase, their forms exhibit rich variations in combination with other components. Through Pearson correlation analysis, the results shown in Table 4 were obtained.

Table 4 Results of Pearson correlation analysis between lip morphology, mouth morphology, and other components’ morphology

From Table 4, significant correlations between different morphological components can be observed. Specifically, square lip shows significant positive correlations with handles-free and egg-shaped belly; round lip shows significant positive correlations with short straight neck, truncated-conical neck, gallbladder-shaped belly, and square columnar belly; flaring mouth shows significant positive correlations with pierced handles, short constricted neck, and lobed belly; foliated mouth shows significant positive correlations with two loop handles, long constricted neck, lobed belly, and spherical belly; trumpet mouth shows significant positive correlations with long cylindrical neck and spherical belly; contracted mouth shows significant positive correlations with short constricted neck and square columnar belly; dish-shaped mouth shows significant positive correlations with two loop handles, phoenix-shaped handle, fish-shaped handle, long constricted neck, long cylindrical neck, angular shoulder, gallbladder-shaped belly, and cylindrical belly; outward-slanting mouth shows significant positive correlations with pierced handles, long cylindrical neck, and sloping shoulder; bowl-shaped mouth shows significant positive correlations with trumpet neck and curved belly; straight mouth shows significant positive correlations with handles-free, short straight neck, short cylindrical neck, and egg-shaped belly.

Notably, dish-shaped mouths show significant positive correlations with multiple morphological features (two loop handles, phoenix-shaped handles, dragon-shaped handles, fish-shaped handles, long constricted neck, long cylindrical neck, angular shoulder, gallbladder-shaped belly, and cylindrical belly). These diverse combinations reflect the widespread application of dish-shaped mouths across different types of vessels. Particularly in Southern Song Dynasty vases, the combination of dish-shaped mouth and gallbladder-shaped belly allowed for multiple uses - not only as flower vases but also as chopstick vases, as shown in Fig. 11. This usage is supported by images of chopstick vases in Southern Song tomb murals: in the kitchen scene on the east wall of the tomb chamber, among 10 people busy with kitchen work, one maid’s tray contains a vase with chopsticks and an inverted bowl, as shown in Fig. 12⁵³. From the image, this vessel features a shallow dish-shaped mouth, long neck, gallbladder-shaped belly, with the belly diameter slightly larger than the mouth diameter.

Fig. 11: Examples of Porcelain Vases with Dish-shaped Mouth and Gallbladder-shaped Belly.

a Vase with elongated cylindrical neck and dish-shaped mouth, Longquan ware, China, Zhejiang, Song dynasty (960–1279). Source: Daderot, Wikimedia Commons, Public Domain. b Vase, Longquan ware, Southern Song Dynasty (1127–1279). Source: Wikimedia Commons, dedicated to the public domain (CC0). c Green-glazed vase, Longquan kiln, Forbidden City Palace Museum. Source: Gary Todd, Wikimedia Commons, dedicated to the public domain (CC0).

Fig. 12

Kitchen Scene, excavated from Southern Song tomb at Tanxi Village, Xiangyang, Collection of Xiangyang Museum. Source: publicly published papers⁵⁰.

Handle morphology not only affects the functionality of Porcelain Vases but also largely determines their overall esthetic appeal. The combinations of handle morphology with other components were analyzed, with results shown in Table 5.

Table 5 Results of Pearson correlation analysis between handle morphology and morphologies of neck, shoulder, belly, and foot

Table 5 shows that in handle-neck combinations, significant positive correlations exist between handles-free and short cylindrical neck, two loop handles and long constricted neck, pierced handles and short constricted/long cylindrical neck, phoenix-shaped handles and long cylindrical neck, and dragon-shaped handles and long cylindrical neck. In handle-shoulder combinations, significant positive correlations exist between two loop handles and full shoulder, pierced handles and sloping shoulder, phoenix-shaped handles and angular shoulder, dragon-shaped handles and angular shoulder, and fish-shaped handles and angular shoulder. In handle-belly combinations, significant positive correlations exist between two loop handles and olive-shaped belly/curved belly, pierced handles and spherical belly, phoenix-shaped handles and cylindrical belly, dragon-shaped handles and cylindrical belly, and fish-shaped handles and cylindrical belly. In handle-foot combinations, significant positive correlations exist between handles-free and hollow ring feet, and between two loop handles and solid ring feet.

These morphological combinations reflect the sophisticated balance between functionality and decorative qualities in Longquan porcelain vases. For example, the combination of handles-free with short cylindrical neck is suitable for daily practical vessels, being simple and convenient for everyday placement and use; while combinations of phoenix-shaped or dragon-shaped handles with long cylindrical neck is commonly found in ceremonial or highly decorative contexts, conveying elegant and dignified ceremonial presence. Such combinations were common in Southern Song Dynasty porcelains, symbolizing aristocratic and scholarly esthetic preferences. In handle-shoulder combinations, pierced handles with sloping shoulder, phoenix-shaped handles with angular shoulder, dragon-shaped handles with angular shoulder, and fish-shaped handles with angular shoulder were common combinations, enhancing vessel stability and overall harmony while facilitating precise control during firing, thus improving production success rates. Additionally, in handle-belly combinations, the flowing lines of cylindrical bellies complement decorative handles such as phoenix-shaped and dragon-shaped handles, enhancing esthetic appeal, as shown in Fig. 13. In handle-foot combinations, handles-free with hollow ring feet and two loop handles with solid ring feet were also common, with the former ensuring light weight and stability during firing, while the latter provided more substantial support for storage vessels.

Fig. 13: Representative examples of Porcelain Vases Showing Handle and Belly combinations.

a Vase with Phoenix Handles, Longquan ware, Southern Song Dynasty (c.1127–1279 C.E.). Source: Anonymous, Wikimedia Commons, licensed under CC BY 4.0. b Porcelain vase, Longquan kilns, Southern Song Dynasty (1127–1279), Zhejiang Provincial Museum. Source: Gary Todd, Wikimedia Commons, dedicated to the public domain (CC0). c Vase with two fish-dragon handles, Longquan kilns, Southern Song or Yuan Dynasty (late 13th to early 14th century). Source: Daderot, Wikimedia Commons, dedicated to the public domain (CC0).

Neck morphology serves a crucial transitional and connecting role between the mouth and belly of Porcelain Vases, with different combinations significantly influencing the vessel’s overall stylistic characteristics. The analysis results are shown in Table 6.

Table 6 Results of Pearson correlation analysis between neck morphology and morphologies of shoulder, belly, and foot

From Table 6, it can be observed that in neck-shoulder combinations, significant positive correlations exist between sloping shoulder and long cylindrical neck, full shoulder and both short cylindrical neck. Thus, there are three types of neck-shoulder combinations: sloping shoulder-long cylindrical neck, and full shoulder-short cylindrical neck/long constricted neck. In neck-belly combinations, among short neck, significant positive correlations exist between short straight neck and egg-shaped belly, short cylindrical neck and curved/cylindrical belly, and short constricted neck and curved belly; among long neck, significant positive correlations exist between trumpet neck and gallbladder-shaped belly, truncated-conical neck and flattened belly, long constricted neck and lobed belly/curved belly, and long cylindrical neck and flattened/spherical belly. Consequently, there are eleven types of neck-belly combinations: short straight neck-egg-shaped belly, short cylindrical neck-curved belly, short cylindrical neck-cylindrical belly, short constricted neck-curved belly, trumpet neck-gallbladder-shaped belly, truncated-conical neck-flattened belly, long constricted neck-lobed belly, long constricted neck-curved belly long cylindrical neck-flattened belly, long cylindrical neck-gallbladder-shaped belly and long cylindrical neck-spherical belly.

Among these, the combination of long constricted neck and lobed belly creates a visual contrast, giving the vessel both elongated lines and rich curvature with a segmented feel. The lobed vase is a typical representative of this style, with specific variations showing distinct characteristics depending on proportions and design details. Additionally, no significant correlations were found between neck and foot morphologies. This indicates high independence in the design of these two components, where the morphological characteristics of one do not significantly influence the design of the other.

The combination of shoulder and belly morphologies shapes the overall profile of Porcelain Vases, while foot morphology determines their stability and support. Table 7 presents the detailed analysis results of combinations among these three components.

Table 7 Results of Pearson correlation analysis between shoulder morphology and morphologies of belly and foot

Table 7 shows significant positive correlations between sloping shoulder and curved belly; between angular shoulder and both olive-shaped and cylindrical belly; full shoulder and curved belly; and between full shoulder and cylindrical belly. This yields five types of morphological feature combinations. The sloping shoulder-curved belly combination creates flowing lines with a rounded, full overall form. The angular shoulder-olive-shaped belly pairing allows the vessel’s midsection to expand outward, creating a compact yet dynamic form that enhances the porcelain vase’s sense of layering. The angular shoulder-cylindrical belly combination pairs clean straight lines with rounded belly contours, producing symmetrical and orderly forms. The full shoulder-cylindrical belly design imparts a sense of solidity to the vessel body, commonly used in larger porcelain vases suitable for display and ceremonial occasions.

Finally, the combinatorial relationships between lip and mouth, and between belly and foot were analyzed, with results shown in Tables 8 and 9. Square lips show significant positive correlations with flaring mouths and straight mouths; round lips show significant positive correlations with contracted mouths. This yields three types of morphological combinations: square lip-flaring mouth, square lip-straight mouth, and round lip-contracted mouth. In belly and foot morphology combinations, olive-shaped bellies show significant positive correlations with solid ring feet.

Table 8 Pearson correlation analysis between lip and mouth morphologies

Table 9 Pearson correlation analysis between belly and feet morphologies

In summary, through Pearson correlation analysis, this study identified 76 significant morphological feature combinations in Longquan Porcelain Vases, as shown in Fig. 14. The figure employs a radial structure to display the combinatorial relationships among seven major components, with line thickness indicating the magnitude of correlation coefficients and different colors distinguishing combinations between different components. As evident from the figure, the connections between mouth-neck and neck-belly are the most dense, reflecting the richest combinatorial relationships among these components; while the foot shows fewer connections with other components, indicating its relatively independent morphology. The dish-shaped mouth forms significant combinations with multiple components, demonstrating its representative status in Longquan porcelain vase design.

Fig. 14

Structural diagram of morphological feature relationships in Song Dynasty Longquan Porcelain Vases.

Discussion

The main findings of this study are threefold. First, the morphological system of Song Dynasty Longquan porcelain vases demonstrates “diversity within standardization,” reflecting ordered variations under strict craft regulations that balance traditional standards with artistic expression. Second, a systematic correspondence exists between morphological features and functional attributes, embodying the “form follows function” principle where practical utility and esthetics achieve unity through thoughtful design. Third, Longquan porcelain vase morphology embodies seemingly contradictory but unified cultural trends unique to the Song Dynasty—commitment to traditional inheritance alongside innovation, and pursuit of simplicity alongside elegance—reflecting both the influence of literati esthetics and the Longquan Kiln’s ability to maintain cultural continuity during technical advancement.

The contributions of this study are multifaceted. First, this research effectively integrates qualitative and quantitative methods in cultural heritage craft research, and by combining systematic coding from grounded theory with Pearson correlation analysis, reveals intrinsic relationships between morphological features. Second, through establishing a multi-level feature classification system, this study validates Thér’s emphasis on combining intuitive observation with precise measurement, deepening the understanding of traditional craft technical systems⁵⁴. Furthermore, this research systematically examines the interactive relationship between morphological characteristics and cultural elements, complementing Wang and Marwick’s findings about artifact standardization levels reflecting changes in social organization⁵⁵, as well as Piliciauskas et al.’s confirmation of the importance of porcelain morphological analysis in understanding cultural exchange⁵⁶, constructing an analytical framework connecting material features with cultural content. Finally, the framework constructed by this research provides a theoretical foundation for digital cultural heritage protection and dissemination; its morphological feature classification system and coding methods could potentially provide references for AI systems in constructing ceramic morphological feature recognition standards that better align with cultural meanings; the morphological combination patterns revealed might provide data references for digital restoration technologies, enabling virtual restorations of damaged ceramics based on verified combination patterns, supporting the adoption of a U-Net–based convolutional neural network architecture for AI-assisted restoration, and underpinning the establishment of an open-access database to share annotated morphological data and restoration models; the systematic analysis of vessel profile curves could potentially provide basic data structures for mathematical analysis methods such as fractal dimensions, contributing to the exploration of geometric patterns and esthetic principles in traditional ceramic shapes, and enabling the application of box-counting and multifractal-spectral algorithms for advanced fractal feature analysis. These explorations promote cross-disciplinary research among archeology, art history, computer science, and other disciplines, and provide methodological references for practical fields such as cultural creative product design, digital museum exhibitions, and art authentication. In conclusion, through comparative analysis with existing research, this study’s innovative value lies in its methodological integration, analytical depth, and cultural interpretation framework. The theoretical significance of these innovations is providing an operable and verifiable scientific method for traditional craft studies. Their application potential is reflected in data support for craft preservation, morphological language references for cultural creative design, and new pathways for digital cultural heritage research.

Although this study has yielded some promising results, it still presents certain limitations that are expected to be addressed and refined in subsequent research. First, reliance on published literature may skew the sample toward well‑preserved, high‑artistic‑value or historically significant ceramics—at the expense of ordinary daily‑use wares—which could lead to an overrepresentation of fine ware characteristics rather than a complete portrait of Song‑era Longquan bottles; future work should therefore broaden the typological spectrum by drawing on first‑hand sources (museum archives, archeological excavation reports and private collections), with particular emphasis on everyday vessels. Second, there are imbalances in the geographical and temporal distribution of samples: geographically, most literature focuses on Longquan ceramics unearthed within China with limited records of overseas collections (which may include variants with special export characteristics); temporally, although both Northern and Southern Song periods are covered, their proportional representation may not reflect actual production circumstances—so subsequent research should strengthen targeted inclusion of overseas holdings and deliberately balance samples from both Song periods through comparative analysis of domestic and international finds to more accurately reveal regional characteristics and diachronic evolution. Third, although grounded‑theory coding was systematized via multiple expert verifications, literature validation and detailed coding rules, subjective factors remain inherent in any visual‑form–based qualitative method; to mitigate this, future studies will integrate more objective technologies—three‑dimensional scanning, computer‑vision analysis and machine‑learning classification—to reduce bias, improve the accuracy and consistency of morphological feature identification, and thus enhance the universality of the findings. Fourth, the lack of interactive visualization options currently limits users’ ability to explore complex, multidimensional morphological relationships; future efforts will therefore focus on a user‑friendly, web‑based platform—offering dynamic filtering, linked views, and real‑time parameter adjustment—to empower researchers and practitioners to interrogate the data more deeply, test hypotheses on the fly, and share bespoke visualizations, thereby extending the study’s reach and practical impact. The continued development and diverse integration of these methodological advances promise to enrich our cognitive depth of the Song‑dynasty ceramic craft system and open broader avenues for traditional craft research.