The Acoustics of Aggregation Sites: Listening to the Rock Art Landscape of Cuevas de la Araña (Spain)

ABSTRACT

Since the 1980s, research into aggregation sites has focused on the material dimension of the archaeological record and has thus led to an incomplete view of the prehistoric reality. Early communities chose sites for many reasons. In this article, we explore the possibility that something as immaterial in nature as the acoustics of the sites may have influenced the selective choices made by the prehistoric peoples who created and used rock art in these meeting places. Employing the Impulse Response (IR) methodology, we systematically analyze the acoustic properties of Cuevas de la Araña, an aggregation site in eastern Spain with Levantine rock art, and 16 satellite sites in its surroundings. The results indicate that the acoustics of La Araña could have been important for creating affective engagement through musical performances carried out during the production and use of rock art, as well as for the choice of the site as a meeting place for Levantine populations.

KEYWORDS:

• Iberian Peninsula
• Mesolithic
• Neolithic
• Levantine rock art
• archaeoacoustics

Introduction

Periodic aggregation for social or material reasons is a common practice in hunter-gatherer societies (Barnard 1999, 58; Shott 2004, 69; Kelly 2013, 174). As far as prehistoric contexts are concerned, several authors consider that, in these large gatherings, human groups that were usually scattered throughout the territory for most of the year would meet at specific places in order to exchange partners and material goods (Hofman 1994, 348; Miller et al. 2019, 47; Kim 2021, 1); explore seasonal resources (Binford 1978, 347); transmit strategic information and beliefs related to collective memory (McDonald and Veth 2006, 110; Fritz, Tosello, and Conkey 2016, 1325; Maher and Conkey 2019, 93); strengthen kinship and solidarity ties through initiation ceremonies, healing acts, and marriages (Conkey 1980, 810; McDonald and Veth 2012, 93; Bourdier 2013, 377); and, undertake activities linked to the production and use of rock art (Bahn 1982, 265; Moure-Romanillo 1994, 322). The last of these activities would have been immersed in ritual contexts consisting of ceremonies, songs, and dances, in which the images would have been understood much more as a means than an end, since they would have been used to visually recall and celebrate the cultural significance and supernatural power of these gathering places (Domingo et al. 2020, 702). These aggregations would also be practiced by agricultural and pastoral populations in the Iberian Peninsula (Fairén 2006) and other parts of the world (Goring-Morris and Belfer-Cohen 2011). Thus, for generations, these sites, charged with social and symbolic significance, would have functioned as anchors, forming cultural landscapes inhabited by groups that had mobility as a fundamental characteristic (Bueno 2009, 346; González-Ruibal and Torres 2018, 23).

Anthropological research into the aggregation and dispersion processes of pre-industrial societies began in the early years of the 20th century a.d. Mauss and Beuchat’s pioneering study (1904–1905) of the seasonal movements of Eskimo populations was of particular importance. However, archaeology’s interest in aggregation sites would only awaken in the 1980s after the publication of Margaret Conkey's classic work on the Altamira Cave (Spain) as a meeting place for Magdalenian hunter-gatherers. Following Conkey’s study (1980), numerous interpretations of prehistoric sites as possible places of aggregation would emerge in Europe. Palaeolithic examples include the French sites of Isturitz and Mas d'Azil (Bahn 1982), Laugerie Basse (Conkey 1992), Roc-aux-Sorcier (Bourdier 2013), and Marsoulas (Maher and Conkey 2019) and the Spanish caves of El Castillo, Candamo, and Santimamiñe (Utrilla 1994). In post-Palaeolithic times, possible periodic meeting places are the Swedish site of Nämforsen (Fuglestvedt 2018) and, in Spain, certain Galician sites with petroglyphs (Santos Estévez 1998, 2007) and various shelters with Levantine rock art in the Mediterranean area of the Iberian Peninsula (Bader 2002; García Atiénzar 2011; Bea 2012). The last of these are addressed in greater detail in this article.

Despite the emphasis placed on aggregation sites during the last four decades, a key issue continues to be the determination of the attributes that allow the archaeological identification of this type of site (Milesi 2018, 22; Vogels, Fäder, and Lenssen-Erz 2021, 151). In this respect, an analysis of the literature devoted to the subject indicates that, for most researchers, the spaces where prehistoric populations would have gathered tend to present—although not necessarily in all cases—the following characteristics: 1) a strategic position in the landscape and dimensions that allow a large number of people to gather¹ (Bahn 1983, 334; Conkey 1992, 22; Utrilla 1994, 98–99; Utrilla and Bea 2008, 116; Bourdier 2013, 377); 2) they are surrounded by smaller, chronologically contemporary satellite sites and are placed approximately 40–60 km away from other aggregation sites (Conkey 1980, 617; Bahn 1983, 335; McDonald and Veth 2012, 97); 3) they have occupation levels containing a large number of artifacts with typological and stylistic diversity, the latter being greater at the aggregation site than at the satellite sites in its surroundings (Conkey 1980, 612; Bahn 1982, 265; Utrilla 1994, 98; Utrilla and Bea 2008, 116); 4) they present material evidence of recurring periodic occupations (Conkey 1980, 614; Bourdier 2013, 377; Maher and Conkey 2019, 115); and, 5) they yield material culture of a symbolic nature, such as personal adornments, rock art, or portable art (Bahn 1982, 265; Guráieb 2001, 376; Utrilla and Bea 2008, 116). With regard to rock art specifically, panels from aggregation sites are also considered to have more figures, a broader repertoire of motifs, and a greater stylistic variability compared to panels from satellite and other non-aggregation sites in the same region (Bea 2012, 288; McDonald and Veth 2012, 97; 2013, 77; Vogels, Fäder, and Lenssen-Erz 2021, 155).

Examining this set of characteristics, it is possible to observe that they clearly reflect the materialist paradigm on which traditional archaeology is structured. Despite being fundamental for the identification of aggregation sites, these attributes are based on Western ontologies that take into account only the tangible and visible aspects of the archaeological sites, although the prioritization of visual perception or materiality is not a general rule of human societies (Beaune 2018, 75). Thus, this materialist perspective does not consider the fact that an intangible aspect, such as acoustics, could also have influenced the decisions made by prehistoric people when choosing places to organize their aggregations and create their artistic representations.

In light of this problem, we pose the following questions: Do the aggregation sites have acoustics that would enhance the perceptual impact of the social and ritual activities carried out in those spaces? Is there a relationship between the acoustic properties of a rock art site and its use as an aggregation site? To answer these questions, we employ the Impulse Response (IR) method to analyze the acoustic properties of Cuevas de la Araña (Spain), a place with Levantine rock art that presents the typical characteristics of an aggregation site, and 16 satellite sites located in its surroundings. Thus, in the following pages, we contextualize Cuevas de la Araña as a meeting place in the middle basin of the Júcar River, revealing the general aspects of its rock art. Then we present the technical details of the methodology used in this study and the results of the acoustic characterization of the shelters. Finally, we propose answers to the two research questions posed at the beginning of this paragraph and discuss the relevance of our data within the framework of archaeological studies of prehistoric aggregation sites.

Cuevas de la Araña: An Aggregation Site in the Middle Basin of the Júcar River

The rock art and its context

Identified in nearly 800 open-air shelters scattered throughout eastern Spain, Levantine rock art is a prehistoric cultural manifestation without parallel on the European continent. Its thematic repertoire includes narrative compositions that depict scenes of hunting, food gathering, dances, conflicts, and other representations that are difficult to interpret (López-Montalvo 2018; Rubio, Viñas, and Santos da Rosa 2019; Santos da Rosa, Fernández-Macías, and Díaz-Andreu 2021). Debates about the chronocultural affiliation of this post-Palaeolithic art have been constant since its discovery in the first decade of the 20th century a.d. (Díaz-Andreu 2012; Ochoa et al. 2020). Currently, the main point of disagreement is on establishing its chronological origin. Some researchers believe that Levantine art emerged during the Epipalaeolithic/Mesolithic (Viñas et al. 2016; Mateo Saura 2019), while others place it in the early Neolithic (Villaverde et al. 2012).

One of the best-known and most emblematic sites of this rock art tradition is the Cuevas de la Araña, located in the municipality of Bicorp (Valencian Country) in the center of the territory corresponding to the middle basin of the Júcar River, an area with 106 sites² with post-Palaeolithic rock art (Martorell Briz 2019) (Figure 1). The site’s placement in the landscape and its topographic location—in the so-called Hongares gorge—presents particularities that distinguish it from the rest of the painted rock shelters in the region. It is classified as a hidden site, according to the model proposed by Martínez (1998).³ It is located some 50 m upstream from where the gorge plunges an imposing 60 m at the so-called Tajo de la Rebolla and narrows into a transversal “V” shape. The width between the limestone banks is 10–20 m at the bottom—where a stream runs, the flow of which can be interrupted during the dry season—and 50–60 m at the top of the slope.

Figure 1. Map of the study area with the location of the rock shelters whose acoustic properties were analyzed.

In terms of its spatial structure, the site is made up of a main shelter with two decorated cavities (in front of which there is a large natural platform), an unpainted shelter known as the cueva grande, and a third shelter—somewhat distanced from the main shelter—where some depictions are also found (Figure 2). Like the cavities with paintings, the cueva grande is oriented towards the southwest. It has a 14 m wide mouth, a depth of between 4 and 5 m, and a height of 3 m. When the site was first investigated, areas with black soil interpreted as evidence of prehistoric structures were identified on its nearest slope. Flint flakes and fragments of polished ophite axes typical of post-Palaeolithic industries were found in these areas (Hernández-Pacheco 1924, 48). In subsequent studies, lithic artifacts attributed to the Upper Palaeolithic and Mesolithic were found on the surface (Aparicio 1976, 157; 1977, 58; 1979, 244), as well as Bronze Age pottery sherds (Enguix 1981, 256). A new excavation revealed a Mesolithic level covered by another from the Early Neolithic that contained Cardial pottery (Pla Ballester 1984, 83).

Figure 2. General view of Cuevas de la Araña (Photo: N. Santos da Rosa).

The main shelter is located a few meters from the cueva grande. The first of its cavities has an opening approximately 4 m wide, 1 m deep, and 3 m high. The second, adjacent to the previous one, is the most important, both for its dimensions—a 7 m opening, a height of 3 m, and a depth of 2 m—and the large number of rock art motifs. Its panels have paintings associated with different stylistic horizons and overlapping and complex compositions. The third shelter is about 12 m from the main shelter. At the height of approximately 3 m, it is 7 m wide, 1 m high, and 1.5 m deep. Thus, the pictorial ensemble of the site is made up of more than 150 motifs (Martínez i Rubio 2011, fig. III.1), almost all of them in the Levantine style and depicting archers, deer, caprids, equids, foxes, and a bull, among others. The few figures in classical schematic style correspond to a small number of zoomorphs and a tectiform, while the only motif attributable to early schematic art is a zigzag (Figure 3).

Figure 3. A) Tracing according to Hernández-Pacheco (1924); B) photograph taken during the acoustic measurement work (Photo: N. Santos da Rosa).

Cuevas de la Araña as an aggregation site

The subject of aggregation sites has been little explored in the framework of Levantine rock art research. Nevertheless, some authors believe there would have been sites of this type at several places in the Mediterranean region of the Iberian Peninsula. They would have formed “sacred territories” composed of a “main sanctuary” surrounded by “secondary sanctuaries” (Alonso and Grimal 1999, 193). In this respect, Bader (2002) maintains that sites such as Torcal de las Bojadillas and Solana de las Covachas (Albacete) could have been used as meeting places. As that author points out, these are sites that have enough space to receive a large number of people, and they present a high number of painted figures and a greater thematic and stylistic variability compared to satellite sites located in the same region. The identification of characteristics similar to those at Levantine sites in the Aragón area led Bea (2012) to classify the Val del Charco del Agua Amarga, La Vacada, Los Chaparros, El Garroso, El Chopo, and Muriecho shelters as possible aggregation sites. In the case of the Muriecho, which has an exceptional scene of a group capturing a live deer (Utrilla and Bea 2005), the depiction of dancers and characters carrying instruments that appear to correspond to some type of aerophone is striking. This suggests that dance and music would have been part of the activities carried out by the human groups that met and painted their depictions in this place (García Benito 2018; Santos Da Rosa et al. 2021).

Analyzing the characteristics of Cuevas de la Araña and the geographical context in which it is found, we identify clear evidence that this site was probably used by Levantine groups as a place of aggregation. As mentioned above, this site occupies a central position in the middle basin of the Júcar River, located in a monumental landscape and at a strategic point on the network of optimal paths used by the prehistoric populations who inhabited this region (Martínez i Rubio and Martorell Briz 2012). Moreover, it has a natural platform approximately 8 m long × 30 m wide covering an area of some 240 m² in front of the main shelter. A large number of people (approximately 311 standing and 266 sitting or squatting)⁴ could have gathered here. Furthermore, within a radius of 25 km of Cuevas de la Araña, there are dozens of satellite sites with Levantine art, 16 of which have been acoustically analyzed in the context of our research. It should also be noted that La Araña is located some 40 km from Cueva de la Vieja (Albacete), another Levantine complex considered by some researchers to be an aggregation site (Alonso and Grimal 1999).

Although limited in number, the archaeological finds, both from the surface and the identified occupation levels, present the typological and stylistic variability expected at a site frequented repeatedly over time. This also suggests that, during aggregations, the Levantine populations would probably camp at the site—where there was enough space—and not in its surroundings. Moreover, as we have already pointed out, rock art is a frequent characteristic of places of aggregation. In this sense, Cuevas de la Araña stands out as the site with the largest number of figures and the widest repertoire of motifs in the entire middle basin of the Júcar. It is also the only one that has representations corresponding to the six morphotypes that make up the regional stylistic sequence of Levantine art (Figure 4). This sequence, proposed by Martínez i Rubio (2011) and based on the typological characteristics of human figures, begins with essentially static motifs of large dimensions that tend towards disproportion and in which the muscular volumes of the legs and arms, the facial features, and the material culture are detailed with precision. As the sequence progresses, the representations become smaller, their disproportion is more pronounced, and more linear forms are adopted on the extremities and the trunk. In this respect, the changes are not limited to the morphology but are also detected in the themes and the concepts of the graphic space. New themes appear, such as scenes of violence or the collection of honey and other products, in which the incorporation of natural cavities in the bedrock as part of the scenic composition stands out. Finally, of note are the overlaps between Levantine figures and Levantine and schematic motifs, which indicate that the site was seen as a significant place in the landscape for many generations of prehistoric painters.

Figure 4. The six morphotypes of the chrono-stylistic sequence of Levantine rock art in the Júcar River basin (adapted from Martínez i Rubio and Martorell Briz 2012, 72).

Methodology used for the Acoustic Characterization of the Shelters

Description of the study sample

The study sample analyzed in this research includes Cuevas de la Araña and 16 satellite sites with Levantine rock art located in its surroundings (Table 1). The latter were selected according to the following parameters: 1) the state of conservation of the original morphological characteristics of the shelter; 2) the accessibility of the site and adequate space for the assembly of the acoustic measurement equipment; and, 3) the state of conservation of the rock art.

Table 1. The rock art sites that make up the study sample: dimensions, number of Levantine motifs, and their chrono-stylistic classification.

Rock art site	Dimensions (width × height × depth)/orientation	Number of Levantine motifs	Chrono-stylistic Levantine sequence
Cuevas de la Araña (CA)	15 × 3 × 2 m/SW	+150	T1, T2, T3, T4, T5, T6
Abrigo de Voro (VO)	30 × 1.5 × 2 m/E	115	T2, T3, T4, T5, T6
Abrigo del Garrofero (AG)	23 × 4.5 × 4.5 m/W-SW	28	T1, T3
Alto Bolinches (AB)	62.5 × 4.3 × 4.5 m/NW	10	T2
Corrales del Tío Alfredo Cosechas I (TA1)	11.25 × 1.77 × 2.25 m/NW	6	T5
Cuevas Largas II (CLa2)	90 × 4 × 4.5 m/E	14	T5
Cuevecicas del Estiércol II (CE2)	205 × 4.5 × 4 m/SE	4	T3, T5
Río Grande II (RG2)	41 × 3.75 × 4.2 m/NW	1	Indeterminate
Abrigo de Lucio o Gavidia (LU)	46 × 5.2 × 3.5 m/NW	100	T3, T4, T5
Balsa de Calicanto (BC)	27 × 3.4 × 4.7 m/SW	3	Indeterminate
Abric de Trini (AT)	8.45 × 1.8 × 1.3 m/E	94	T4
Barranco del Randero II (BR2)	16.8 × 4.4 × 4 m/W	10	T2
Abrigo de las Cañas I (Ca1)	5.2 × 3.6 × 3 m/S-SW	18	T2
Abrigo de los Chorradores (Cho)	8.5 × 2.4 × 2.7 m/N	24	T5
Cinto de las Letras (CL)	12 × 5.5 × 3 m/SE	91	T1, T2, T3, T4, T5
Abrigo de Jesús Galdón (JG)	20 × 4 × 2.55 m/W	2	Indeterminate

Regarding the first parameter, some of the satellite shelters in the study area have lost their original morphological characteristics due to the construction of walls to use them as cattle pens or to restrict access to the site for conservation reasons. This makes it impossible to carry out experimental archaeoacoustic studies in these places. Regarding the second parameter, it is currently not possible to safely access some of the sites in the region, as the shelters may be at a height of several meters and/or in places where the access path has collapsed. In other cases, although it is possible to access the site, its small size does not allow the assembly or correct operation of the acoustic measurement equipment. This must be positioned according to certain criteria: for example, the microphones must be at least 1 m from the shelter walls or any other obstacle. Finally, regarding the third parameter, the state of conservation of the rock art at some sites makes it impossible to determine whether the painted images correspond to Levantine rock art representations. Therefore, we did not include them in our sample.

Description of the measurement method

The acoustic characterization of the rock art shelters was based on the analysis of a set of impulse responses (IR) recorded at each site. Generally speaking, an impulse response is a signal that describes how a space modifies a sound reproduced at a particular emission point prior to its arrival at the ears of a person located in a specific receiver location. Thus, the measurement method followed to gather IR in the rock art shelters of the study sample was carried out following the procedure established by the International Organization for Standardization (ISO 3382-1 2009), while considering the specific needs and limitations of outdoor spaces. In this respect, it is worth mentioning that, despite being designed for indoor performance spaces (theaters, auditoria, concert halls, etc.), measurement protocols based on the aforementioned standard have been successfully applied in a great variety of other places, such as glacier caves (Malecki et al. 2020), open-air theaters (Farina and Tronchin 2013; Hak et al. 2016; Astolfi et al. 2020), and open-air shelters with rock art (Rainio et al. 2014; Mattioli et al. 2017).

Within the scope of this study, a representative set of IRs was collected at each of the sites to analyze their acoustic properties, including a minimum of two source (S) positions and two receiver/listener (R) positions. Both S and R locations were selected taking into account the size and morphology of each shelter and the possible use prehistoric societies may have made of it. Additionally, for some of the S-R combinations, 2–3 IRs were recorded consecutively to evaluate the repeatability of the results.

The IR measurements were based on the exponential sine-sweep technique (ESS) developed by Farina (2000). The duration, frequency range, and emission level of the exponential sine-sweep used as an excitation signal were determined to ensure the sufficient quality of the recorded IR, taking into account both the environmental conditions and the limitations of the equipment. Specifically, the excitation signal used to record the IR included in this study was a 12 second long exponential sine-sweep from 50 Hz–20 kHz. The emission level was adjusted for each site to ensure a minimum impulse-to-noise ratio (INR) of 45 dB⁵ (Hak, Wenmaekers, and Luxemburg 2012). The INR (dB) is an indicator of the IR decay range and, therefore, of its quality.

The equipment used for the measurements met the ISO 3382-1 standard requirement. It consisted of an omnidirectional sound source (IAG DD4 mini dodecahedral loudspeaker with an IAG AP4GB power amplifier) with a frequency range of 50 Hz–15 kHz that was used to emit the excitation signal. At each receiver point, both monaural and spatial IR were recorded by using an omnidirectional microphone (micW n201) together with an audio device (Zoom F4) and a high-quality 3rd order Ambisonics microphone (Zylia ZM). However, only the results obtained from the recordings with the monaural omnidirectional IR (W) synchronously recorded by the EASERA 1.2 software tool were considered in this study, while the spatial IRs were archived for future analysis and auralization purposes.

EASERA 1.2 was also used to analyze the IRs measured at each receiver position in order to quantify how early reflections coming from the shelter influence sound transmission and to detect if there was any relevant late reflection generated by the surrounding space. Moreover, the software was employed to derive a number of monaural acoustic paraments included in the ISO 3382-1 and related to different listening aspects. Although they were originally used in the study of indoor spaces, where reverberant energy has a significant contribution and diffuse conditions are expected, the use of these parameters in the analysis of open spaces provides important insights into the acoustics of the places and makes it possible to carry out an objective comparison between different sites. In this sense, the two first parameters are early decay time, EDT (s), and reverberation time, T₂₀ (s), both related to the decay of sound in the space, the first being an indicator of the perceived reverberance and the second being more physically important. Clarity or early-to-late energy sound index, C₈₀ (dB), a parameter related to the perceived clarity of sound, mainly used for music, was also gauged. The fourth parameter is the definition, D₅₀ (-), another early-to-late energy ratio usually used to assess speech definition or clarity. Finally, we assessed sound strength, G (dB), a parameter related to the subjective sound level and proven to be critical when evaluating the acoustic quality of a space (Bradley 2011).

All the parameters mentioned above are calculated from the omnidirectional IR, while the reckoning of G needs an additional reference signal. Thus, an in situ calibration method was used to estimate the reference signal (Katz 2015). Despite not being detailed in the ISO standard and presenting some deviations compared to the laboratory calibration related to signal processing (Katz 2015; Wenmaekers and Hak 2015), this method is broadly accepted in the scientific community and has been used in previous studies with reliable results (Astolfi et al. 2020). The method consists of registering an IR close to the sound source (ideally at 1 m) to avoid very early reflections and isolating the direct sound by windowing the IR over a determined time interval (typically 3–7 ms). At the rock art sites in our study sample, an IR registered at max S-R distance of 5 m was windowed with a time window of 5 ms, allowing accurate results starting from the 250 Hz octave band (although the most reliable results are found in the 500–1 kHz bands).

It may be worth mentioning that parameters related to spatial impressions, such as the interaural cross-correlation, IACC (-), and early lateral energy fraction, J_LF (-), were not considered in this study, as they are significantly dependent on both the positioning and orientation of the microphones. This dependency complicates drawing general conclusions in open spaces, especially from a low number of measurements, and makes it difficult to compare the results between the rock art sites due to their variability in shape and size.

Following the standard, both IRs and acoustic parameter values were analyzed in the octave bands from 125 Hz–4 kHz. Considering an extended range from 63 Hz–8 kHz would have been desirable, since those frequencies could be relevant in terms of perception (Bradley 2005), but it was not possible due to technical limitations inherent to the equipment required to perform the measurements. In this respect, the small size of the dodecahedral loudspeaker—crucial for its portability—limits its performance in terms of power at low frequencies (under 50–100 Hz). Therefore, the 63 Hz octave band is not considered, and the results with the 125 Hz band need to be analyzed with caution. Furthermore, the directivity pattern of any dodecahedral loudspeaker is not actually omnidirectional over a certain frequency. In this case, the directivity of the sound source used would affect the results above the 4 kHz band and, for that reason, the 8 kHz frequency band is not included in the analysis.

Regarding data usability, the spectral and spatial analyses of the results were used to examine the acoustic properties of each site individually, so that detailed information was obtained regarding the variation of their acoustics depending on the frequency and/or position considered. Mid values (single values’ frequency averaged according to ISO 3382-1) of the acoustic parameters obtained at a reference position were used to compare the selected sites. The reference position is defined as the one with the sound source placed in front of the main rock art panel (or the center of the shelter if it is entirely covered with paintings) and the receiver positioned about 5 m from the sound source.

Results of the Acoustic Measurements

Cuevas de la Araña acoustic properties

Six tests were performed to characterize the acoustic properties of the Cuevas de la Araña rock art site. They involved three sound source positions located inside and in front of the main shelter and three microphone positions distributed throughout the same area (Figure 5). The INR obtained at each IR greatly exceeded the 45 dB criteria in all the frequency bands of interest, confirming that the recordings had the dynamic range required for a proper calculation of the acoustic parameters (Hak et al. 2016). In addition, the environmental conditions were monitored during the measurement session. We registered an ambient temperature of between 21.5 and 22.8 °C, a mean wind speed of 0 m/s with gusts of 1.5–2.3 m/s, and a background noise level of a L_Aeq = 28.80 dB (low noise environment). The background noise level was measured with the ARTA SPL Meter (Mateljan 2019) as an A-weighted equivalent sound level over a period of 2.5 minutes.

Figure 5. Sound source and receiver positions used to analyze the Cuevas de la Araña acoustic properties (Photo: N. Santos da Rosa).

Analyzing the spectra presented in Figure 6, we can observe that the values obtained for C₈₀ (around 15 dB) and D₅₀ (close to 1) are within the norm for an open-air shelter and denote excellent sound clarity for music and speech. Surprisingly, the T₂₀ of 0.9 seconds is much higher than expected for an open space, corresponding to a value typically found in medium-sized enclosed spaces, such as conference rooms, lecture halls, and cinemas (Beranek 2004). For this reason, the IRs and their Schroeder integrals (Schroeder 1965) are examined in detail. In Figure 7B, the presence of double slope decays, identified in all tests performed at Cuevas de la Araña, can be discerned. The first slope seems to be the contribution of the early reflections of sound coming from the shelter itself, while the second is the result of later reflections from the canyon (see the delay of the second group of reflections in the echogram shown in Figure 7C). The strong influence of direct sound—a consequence of the short distance between the source and the receiver (5.45 m)—can also be seen as a vertical drop from 0 to 8 dB in the Schroeder integral. The non-linearity of the decay curves might lead to an inaccurate calculation of the reverberation time: therefore, T₂₀ values must be interpreted with caution. It is clear that the landscape in which the Cuevas de la Araña is located creates a certain reverberation level, a priori unexpected, as it is an open space. In this regard, the low EDT values, around 20 ms, can be explained by examining the first decays of the IR recorded at the site. Even so, the average T₂₀ of 0.9 seconds suggest that the shelter was a suitable place not only for speech transmission, but also for the production and appreciation of music (Harris 1991), especially when considering the lack of reverberation inherent to the open conditions of the other rock art sites in the study area. Moreover, no echoes can be discerned in any of the IR according to the Dietsch and Kraak (1986) echo criterion, given that the curve remains well below the brown line in Figure 7D.

Figure 6. Spectral behavior of the acoustic parameter values related to A) reverberation perception (EDT and T₂₀), B) music clarity (C₈₀), and C) definition (D₅₀), spatially averaged to include all the tests used for the acoustic characterization of Cuevas de la Araña. As can be seen, all the parameters show an almost flat profile in frequency, which means there is no significant variation for any of the parameters depending on the frequency band considered (error bars show the standard deviation). Reverberation parameters were estimated from double-slope decay curves.

Figure 7. A) Impulse response measured in Cuevas de la Araña at the receiver position R02 when the source was located at S01; B) Schroeder integral (Schroeder 1965); C) echogram; and, D) echo speech function calculated according to Dietsch and Kraak echo criterion (Dietsch and Kraak 1986).

Regarding the spatial behavior of the acoustic parameters, it is possible to observe in Table 2 that the distance between the sound source and the receiver is shorter when both are in the shelter (S01-R02 and S02-R01). Despite this, the EDT values are around 12 ms higher than those identified in tests involving S/R positions placed on the natural platform, probably due to the very early reflections coming from the shelter. Thus, considering that the just noticeable difference (JND) for this parameter is set at 5% (ISO 3382-1), which in this particular case would correspond to only 1 ms⁵, the subjective perception of reverberance is higher when the sound source and the listeners are located inside the shelter. Furthermore, it should be noted that, although energy parameters such as C₈₀ and D₅₀ are highly susceptible to spatial variation, in Cuevas de la Araña, they were not affected by changes in the sound source and receiver positioning. However, in the case of sound strength, G, a level of around 4 dB higher than that expected in free-field conditions using the same measuring system is found when the source or the receiver is located on the platform. Since the JND for G is set as 1 dB, this means that there is a significant natural amplification of sound caused by the shelter and the gorge at those positions. In turn, when the source and the receiver are placed in the shelter, the amplification is less noticeable, showing a gap smaller than 2 dB in relation to the free-field linear regression (Figure 8).

Figure 8. Sound strength mid-frequency averaged (G_m) values calculated for each test as a function of S-R distance (in logarithmic scale). The free-field linear regression (red line) is also reported as a reference.

Table 2. Acoustic parameters single values obtained for each acoustic test performed at Cuevas de la Araña. The frequency values were calculated according to the ISO 3382-1 standard, averaging the values recorded at 500 Hz and 1 kHz frequency bands.

Source ID	S01		S02		S03		Spatial averaged
Receiver ID	R02	S01-R03	R01	R03	R01	R02
S-R distance (m)	5.45	8.08	5.45	7.80	8.08	7.80
$\mathrm{E}\mathrm{D}{\mathrm{T}}_{\mathrm{m}}\left(\mathrm{s}\right)$*	0.21	0.12	0.27	0.13	0.14	0.13	0.17
$T_{30m}\left(\mathrm{s}\right)$*	0.99	0.88	0.92	0.93	0.94	0.83	0.92
Gm (dB)	6.75	6.10	7.10	6.20	5.10	6.80	6.34
$C_{80m}\left(\mathrm{d}\mathrm{B}\right)$	13.55	16.50	13.55	15.40	14.50	16.25	14.96
$D_{50m}$ (-)	0.95	0.97	0.95	0.96	0.96	0.97	0.96

*Reverberation parameters were estimated from double-slope decay curves.

Acoustic properties of the satellite sites

The acoustic properties of the 16 satellite sites were measured and analyzed following the same procedures used in the study of Cuevas de la Araña. Thus, tests were performed according to similar source and receiver positions, taking into account the morphology of the shelter and the location of the rock art panels. Moreover, the monitoring of environmental conditions registered an average temperature of 26.5°C, a mean wind speed of 0 m/s with sporadic gusts of up to 2.9 m/s, and a background noise of a L_Aeq < 30 dB, i.e. similar conditions to those recorded for Cuevas de la Araña.

Figure 9 shows the values for T₂₀, EDT, G, C₈₀, and D₅₀, spectrally averaged according to the ISO 3382-1.⁶ We can observe that reverberation conditions were similar at all the sites, exhibiting a T₂₀ of 0.30 seconds and an EDT close to 0.20 seconds. Such low T₂₀ and EDT values indicate that there is no proper reverberant tail at the satellite sites and that those shelters can be considered “dry” in terms of sound. Nevertheless, this absence of reverberation associated with strong early reflections coming from the shelters resulted in high values of sound clarity for both music (C₈₀ well above 5 dB) and speech (D₅₀ values close to 1).

Figure 9. Acoustic parameter values (ISO averaged) derived from the IR measured at the selected reference point of each site: A) reverberation perception (EDT and T₂₀), B) strength (G_m), C) music clarity (C₈₀), and D) definition (D₅₀). The striped bars represent those sites where the S-R distance is shorter than 5 m due to space restrictions. Gray dotted lines mark the reference values. Reverberation parameters in Cuevas de la Araña (CA) were estimated from double-slope decay curves. The abbreviations used in the figure are listed in Table 1.

The sound strength parameter is also included to assess the natural amplification of sound that each site might cause. As the S-R distance of the tests used for the comparison is 5 m (with a few exceptions due to space restrictions), a G of 6 dB would denote no natural amplification (Table 3). Since the JND set for this acoustic parameter is 1 dB (ISO 3382-1), we can observe that at the VO, AG, AB, BR2, and JG sites (for names, see Table 1), this natural amplification would be clearly perceived, as the obtained G_m is above 9 dB in those places. In that respect, the high G_m value obtained at the BC site is especially remarkable, given that it is more than double the expected value in free-field conditions.

Table 3. Sound strength mid-frequency averaged (G_m) values calculated for each test as a function of S-R distance (in logarithmic scale). The free-field linear regression is also reported as a reference. Reverberation parameters in Cuevas de la Araña (CA) were estimated from double-slope decay curves. The abbreviations used in this table are listed in Table 1.

Site ID	S-R distance (m)	EDTm (s)	T20m (s)	Gm (dB)	C80m (dB)	D50m (-)
CA	5.45	0.21	0.99	6.75	13.55	0.95
VO	5.00	0.15	0.20	10.00	25.60	0.99
AG	5.00	0.07	0.27	9.30	23.55	0.98
AB	5.00	0.08	0.25	10.40	25.45	0.99
TA1	4.20	0.08	0.11	13.00	33.15	1.00
CLa2	5.00	0.09	0.32	8.30	22.35	0.98
CE2	5.00	0.22	0.21	8.55	23.55	0.97
RG2	5.00	0.16	0.26	8.40	22.60	0.98
LU	5.00	0.11	0.35	7.80	20.65	0.98
BC	5.00	0.21	0.32	16.85	18.15	0.94
AT	3.75	0.08	0.10	11.85	35.65	1.00
BR2	5.00	0.15	0.18	9.25	27.60	0.99
Ca1	3.25	0.07	0.09	13.95	34.55	1.00
Cho	4.00	0.09	0.08	13.80	38.30	1.00
CL	5.00	0.16	0.17	8.65	26.15	0.99
JG	5.00	0.15	0.19	9.80	26.90	0.98

Discussion

We return to the first question raised in this article: do aggregation sites have acoustics that would enhance the perceptual impact of the social and ritual activities carried out in these spaces? The results presented in the previous section point to an affirmative answer in the case of Cuevas de la Araña. Thus, the fact that the acoustics of the site are particularly suitable for the production and appreciation of music is highly significant when we consider the possible circumstances of rock art creation at this place. According to ethnoarchaeological studies carried out in Australia, rock art creation at aggregation sites tends to be immersed in ritual contexts in which ceremonies, songs, and dances take place and in which rock art motifs are usually used as mnemonic devices to celebrate the symbolic relevance of these gathering places (Domingo et al. 2020, 702). Musical instruments found in Upper Palaeolithic aggregation sites allow us to consider that this view could also be applied to European prehistory (for a synthesis of these findings, see Jiménez Pasalodos et al. 2021). In Marsoulas, for example, a seashell horn of Magdalenian chronology was used as an aerophone and painted with red dots remarkably similar to the rock art motifs found on the cave walls, which suggests that, in this site, music was part of the activities related to the production and/or use of the images (Fritz et al. 2021; Kolar, Fritz, and Tosello 2022).

Based on this perspective, in prehistoric aggregation sites, rock art could have assumed the role of an active component in performances through which concepts and ontologies would be shared and reinforced among members of society (Feruglio et al. 2019). In this sense, the images would only have wholly fulfilled their function in relation to the environment in which they were located—from which they were inseparable (Lenssen-Erz 2008; Jones 2017)—and its sensitive aspects, such as light, texture, colors, and sound (Zawadzka 2019). Regarding this last element, the importance of sound lies not only in its use as a communication tool but also in its ability to create affective engagement at an inter- and intra-group level (Blake and Cross 2015). This engagement occurs mainly through music, a key element in ritual contexts (Lewis-Williams 1994; Morley 2009) that is capable of influencing human emotions, positively or negatively, in both Western and non-Western societies (Blood and Zatorre 2001; Zatorre, Chen, and Penhune 2007; Juslin and Västfjäll 2008; Thompson and Quinto 2011; Pätynen and Lokki 2016).

In this respect, certain specific aspects of the acoustic properties of Cuevas de la Araña stand out. At this site, the perception of reverberation—an effect through which strong emotional responses can be generated during the appreciation of musical performances (Västfjäll, Larsson, and Kleiner 2002)—is more intense when the sound source and the receiver are inside the main shelter, i.e. in front of the rock art panels and approximately 1 m from the paintings. The relevance of these data stands out when we take into account that, as has been pointed out in ethnographic accounts, in certain cases the very act of creating the images would have been accompanied by the artists singing songs that connected them with the spiritual world (Layton 1992, 59; Díaz-Andreu, Mattioli, and Rainsbury 2021, 246–247). On the other hand, the sound amplification of 4 dB over free-field conditions, identified when the emitter is in front of the paintings and the receiver on the natural platform—where high levels of music clarity and word definition are also observed—indicates that the sounds produced by one or several people positioned in front of the representations could be clearly heard and understood by a large audience. Thus, we can suggest that the acoustics at Cuevas de la Araña would have intensified the sensory effect and emotional impact of ceremonies performed with musical accompaniment when Levantine groups met at the site. This would have facilitated cohesion between the peoples gathered there through affective engagement and would have contributed to alleviating the tensions and scalar stress typical of this type of periodic aggregation (Lee 1972; Alberti 2014).

These data also help us answer the second question posed in our research: is there a relationship between the acoustic properties of a rock art site and its use as an aggregation site? Comparing the IRs obtained at Cuevas de la Araña with those recorded at the 16 satellite sites, we observe some relevant differences. Although the levels of musical clarity and word definition identified at the satellite sites are also very high, the short reverberation time recorded at those shelters makes them suitable for speech transmission but “dry” and lifeless for musical performances (Ando, Okura, and Yuasa 1982; Harris 1991). On the other hand, it should be noted that noticeable acoustic amplification is observed at the Voro, Garrofero, Alto Bolinches, Randero II, Balsa de Calicanto, and Jesús Galdón shelters. As such, considering that most of the satellite rock shelters have morphological and spatial characteristics similar to those of Cuevas de la Araña, the acoustic properties of this site—among other factors—seems to have been decisive both for its use as a place of aggregation as well as for the abundant production of Levantine rock art at the site over a wide spectrum of time.

In this context, we observe an important relationship between our data and the results obtained by Margarita Díaz-Andreu and Carlos García Benito in other regions of the Levantine territory (Díaz-Andreu and García Benito 2015). Although with a less sophisticated methodology than the one used in this study, those authors analyzed the reverberation levels and recorded the presence of echoes at a significant number of rock shelters in the Barranco de la Valltorta (Castellón), the Barranco Mortero (Teruel), and the Sierra de Godall (Tarragona). They aimed to verify whether the Levantine artists had taken into account the acoustic characteristics of the sites when choosing the places in which to paint their representations. In general terms, the researchers concluded that in their three study areas, the sites with the greatest number of figures are precisely those with the highest levels of reverberation and echoes (Díaz-Andreu and García Benito 2015, 59). They particularly noted that in Valltorta, the most notable results were obtained from the La Saltadora, Cavalls, and Civil (Castellón) shelters, considered by those authors as possible places of aggregation (Díaz-Andreu and García Benito 2012, 3597).

The significance of the results obtained in Cuevas de la Araña is also corroborated by the data from recent archaeoacoustic studies carried out in sites of high symbolic relevance related to other rock art traditions located in the Iberian Peninsula and different parts of the world. In northern Spain, Rupert Till, Bruno Fazenda, and their colleagues analyzed the relationship between the location of rock art and the acoustic properties of five Palaeolithic caves: La Garma, Las Chimeneas, La Pasiega, and El Castillo (Cantabria) and Tito Bustillo (Asturias) (Till 2014; Fazenda et al. 2017). After an exhaustive statistical study, these authors found a subtle—although sufficient—correlation between the acoustics of the spaces and the location of the art (Fazenda et al. 2017, 1347) and considered these data as evidence that sound would have been a fundamental element of rituals performed in those places (Till 2014, 299). Also in Spain, Gabriel García Atiénzar and others found that the acoustics of Pla de Petracos—the most representative site of macroschematic rock art—would have been an important element in the scope of social activities and ceremonies carried out to strengthen tribal ties (García Atiénzar et al. 2022, 14). In Finland, Riitta Rainio and her colleagues examined acoustic phenomena recorded at post-Palaeolithic painted sites mentioned in the ethnography of the Sámi people as sacred offering places (Rainio et al. 2014, 2018). In these areas, researchers identified the presence of echoes and observed that the cliffs chosen as a support for the images work as potent sound reflectors. In a specific site located next to Värikallio Lake, for example, the sound seems to emanate directly from the painted panel (Rainio et al. 2018, 543). More recently, Margarita Díaz-Andreu and her research team found that the perceived loudness resulting from a natural amplification of sound, in addition to high levels of music and speech clarity, may have been one of the reasons behind the selection of several locations for rock art production in the Russian Altai (Díaz-Andreu et al. 2022).

Conclusion

Over the last forty years, the subject of aggregation sites has been key to discussions about the social organization and symbolic behavior of prehistoric societies. However, until now, the parameters used to characterize this type of site have focused solely on the tangible dimensions of the archaeological record. Seeking to go beyond these traditional perspectives, in this article, we have investigated whether the acoustics of Cuevas de la Araña could have enhanced the sensory impact of the social and ritual activities carried out there and have influenced the choice of this site as a place of aggregation. Analyzing the sound of the site and comparing the results with those obtained at several satellite sites located in the same region, we find that Cuevas de la Araña presents acoustic characteristics conducive to the realization and appreciation of musical performances that, in ceremonial contexts associated with the production and use of rock art, could create an affective engagement between Levantine groups. Given that these characteristics are not found at the other sites in the study sample, we suggest that the acoustics of Cuevas de la Araña may have been an important factor in choosing it as a meeting place and for the creation of the rock art that can be seen there today.

In a future stage of the research, it will be necessary to ascertain whether these results identified in the middle basin of the Júcar River are also found at other meeting places distributed throughout the different areas of the Levantine territory. However, the evidence presented in this study is already sufficient to corroborate the idea that a suitable archaeological characterization of aggregation sites, such as the one presented in this article, requires the study of their acoustic properties. In this respect, transcending the exclusively materialistic perspective of traditional archaeology to include the intangible aspects of prehistoric reality means accessing a complex network of ontological relationships established by human societies with the landscape and the sensitive dimensions of their environment, a factor that is increasingly fundamental to our reaching a coherent and comprehensive understanding of the past.

Disclosure Statement

The authors report there are no competing interests to declare.

Correction Statement

This article has been republished with minor changes. These changes do not impact the academic content of the article.

Additional information

Funding

This study forms part of the ERC Advanced Grant Artsoundscapes project “The sound of special places: exploring rock art soundscapes and the sacred” (EC Grant Agreement 787842), whose PI is Margarita Díaz-Andreu.

Notes on contributors

Neemias Santos da Rosa

Neemias Santos da Rosa (Ph.D. 2019, Rovira i Virgili University, Spain) is an archaeologist and a postdoctoral researcher at the ERC Artsoundscapes Project—Universitat de Barcelona. Throughout his professional career, he has been researching the technological, social, and symbolic aspects of post-Palaeolithic rock art of the Iberian Peninsula.

Lidia Álvarez Morales

Lidia Álvarez Morales (Ph.D. 2016, University of Seville, Spain) is an acoustic engineer and a postdoctoral researcher at the ERC Artsoundscapes Project—Universitat de Barcelona. She has extensive experience in the architectural acoustics of heritage sites, having worked in multidisciplinary projects developed in Spain, Italy, and England.

Ximo Martorell Briz

Ximo Martorell Briz (Ph.D. 2019, University of Alicante, Spain) is an archaeologist and an honorary research associate at the Prehistory Section of the Universitat d'Alacant/Alacant. His research has been centered on the post-Palaeolithic rock art of the Mediterranean Basin of the Iberian Peninsula.

Laura Fernández Macías

Laura Fernández Macías (B.A. 2011, Autonomous University of Barcelona, Spain) is an archaeologist and a research assistant at the ERC Artsoundscapes Project—Universitat de Barcelona. Her work focuses on the post-Palaeolithic art of the Iberian Peninsula, linking landscape, gender, and acoustics.

Margarita Díaz-Andreu

Margarita Díaz-Andreu (Ph.D. 1990, Complutense University of Madrid, Spain) is an ICREA Research Professor. She is archaeologist and the Principal Investigator of the ERC Artsoundscapes Project. She is a specialist in rock art and archaeoacoustics with extensive international experience, conducting fieldwork in different parts of the world.

Notes

1 Most authors do not specify what they consider a “large number of people.” However, for Conkey (1992, 22), an aggregation site must have enough space to accommodate a group of approximately 50 individuals.

2 Of the 106 sites with post-Palaeolithic art identified in the area, 48 have Levantine art (45.29%) and 36 schematic art (33.96%), while in another 22 (20.75%), both artistic horizons are attested.

3 According to Martínez’s definition (1998, 552), hidden shelters are sites found in ravines of a monumental nature, with high vertical walls below which a river runs. As this author highlights, the fundamental characteristic of these shelters is their restricted visibility at certain points, making them exclusively perceptible from very close by.

4 Adopting parameters previously used by Pastoors and Weniger (2011) and Ruiz-Redondo (2014), Intxaurbe and colleagues (2022) propose that a person would occupy an area of 0.77 m² standing and 0.90 m² sitting or squatting.

5 It should be taken into account that this value, despite having been calculated following the 5% criterion set in the standard, is extremely low, even falling within the precision range of the measuring equipment. Thus, the difference of only one JND is not considered significant. The difference is relevant in this case because it is around 12 JND.

6 Due to the particular characteristics of the open-air shelters, it is not presumed that the results of the acoustic parameters are in the typical ISO 3382-1 range. However, previous knowledge acquired during decades of research regarding the behavior of these parameters in concert halls and other enclosed spaces can be used as a reference.