The term handover describes a joint action between two actors in which an object is transferred from one person to another (Kopnarski et al. 2023b). Handover actions are part of people’s everyday life and are usually performed without requiring much conscious attention. Yet unconsciously, each actor must plan and execute different components of the joint action, such as reaching, grasping, lifting, and transporting the object. In addition, joint action coordination is facilitated through anticipation. For example, the handover position and time must be anticipated, most likely using the giver’s velocity profile during object transport (Mason and MacKenzie 2005). Likewise, the anticipation of object properties (such as its weight) might be integrated into one’s own action plan for grasping the object (Kopnarski et al. 2023a). In other words, the receiver may use the information (implicitly) obtained by observing the giver’s kinematics to adjust their own grip force scaling.

“Joint action can be regarded as any form of social interaction whereby two or more individuals coordinate their actions in space and time to bring about a change in the environment” (Sebanz et al. 2006, p. 70). The success of joint actions depends on, among other things, (i) shared representations, (ii) anticipation of co-actor behavior, and (iii) continuous integration of the anticipated and the monitored information (Sebanz et al. 2006; Sebanz and Knoblich 2021). The term ‘shared representations’ is used to describe when, in a joint action, an actor not only plans their own action execution but creates a mental representation of the co-actor’s action as if they were executing the actions of the other (Kourtis et al. 2014; Schmitz et al. 2017, 2018; Sebanz et al. 2003; Sebanz and Knoblich 2021; Vesper et al. 2013). Regarding the shared representation in a handover action, the receiver forms (i) a mental representation of the giver’s actions (reaching to the object, grasping and moving it toward the receiver, and releasing grip force) (Becchio et al. 2008; Cini et al. 2019; Gonzalez et al. 2011; Meyer et al. 2013). Based on this representation, (ii) anticipations are made regarding the giver’s behavior (Cini et al. 2019; Controzzi et al. 2018; Gonzalez et al. 2011; Huber et al. 2013; Mason and Mackenzie 2005). By monitoring the movement of the giver, (iii) a continuous comparison is made between the anticipated and the observed behavior (Cini et al. 2019; Controzzi et al. 2018; Huber et al. 2013). If there are discrepancies between the anticipated and observed behavior, the receiver’s action plan needs to be adjusted. This comparison between anticipation and observation may provide information about object properties that are not available until object lift-off (e.g., weight). For example, if the receiver anticipates the giver’s motor execution as the giver grasps and lifts the object, a deviation from this anticipation (e.g., an unexpectedly fast object lift) may lead to the inference that the object weight is lower than anticipated.

Joint handover actions can be divided into consecutive phases: (1) reach and grasp (begins when the giver reaches for the object), (2) object transport (begins when the object lifts off), (3) object transfer (begins when the receiver makes initial contact with the object), and (4) end of handover (begins when the giver loses contact with the object) (Kopnarski et al. 2023b). The first two phases have already been intensively researched at the individual action level. Thus, it is already known how various object properties (e.g., size, weight) affect the movement of the actors during object grasping and lifting (Flanagan et al. 2006). Movement trajectories that may provide information about the object’s weight are shaped by the grip and load forces produced when objects are grasped and lifted because the required grip and load forces are primarily determined by the object’s properties, in particular its weight (Hermsdörfer 2009; Schneider and Hermsdörfer 2016; Wing 1996). The heavier the object, the greater the grip and load forces that are required to overcome gravity. Dexterous and efficient grasping and lifting of objects is facilitated through assumptions about the object’s weight and the anticipatory control of the grip forces (Flanagan et al. 2009; Flanagan and Wing 1997b; Johansson and Westling 1988). When the object weight is known, the grip and load forces are anticipated so that the grip force can be adapted to the object weight with heavier objects leading to a higher peak grip force rate (Gordon et al. 1991a; Hermsdörfer et al. 2011; Johansson and Westling 1988). Several studies have indicated that not only prior experience (Flanagan et al. 2008, 2009; Gordon et al. 1993), but also visual cues, such as the object’s size, inform weight anticipation (Gordon et al. 1991a, c; Hermsdörfer et al. 2011; Johansson and Westling 1988). If the object is heavier than expected, the grip force rate is scaled erroneously and has to be increased before the object can be lifted. When assessing whether and, if so, to what extent anticipation matches object weight, the peak of the grip force rate was identified as the most reliable parameter (Hermsdörfer et al. 2011). In contrast to the other force parameters examined, it showed no change when the object weight was manipulated without the participant’s knowledge (Gordon et al. 1991a, c; Hermsdörfer et al. 2011; Johansson and Westling 1988; Li et al. 2009; Nowak et al. 2007). When an object is heavier than expected, the load and grip forces have to be adjusted after initial object contact, i.e., the time from initial contact between fingers and object until the object lift-off is extended: This is referred to as lift delay. When an object is lighter than expected, the initially excessive grip force rate is reduced and the excessive load force results in an earlier object lift-off, i.e., a shortened lift delay (Hermsdörfer et al. 2011; Johansson and Westling 1988; Weir et al. 1991).

Another factor that is influenced by a mismatch between expectation and object weight, is the maximum lift velocity, i.e., the greater the object weight, the lower the lift velocity (Johansson and Westling 1988; Kopnarski et al. 2023a). This relationship between object weight and the grip and load forces not only applies in pure lifting tasks but also in joint actions such as in handover tasks. Both, givers and receivers must produce adequate grip and load forces that correspond to the object weight. First, the giver, who grasps and lifts the object in both the reach and grasp phase and the transport phases, needs to overcome the gravitational force acting on the object. As described above, inaccurate load forces can be detected through a prolonged or shortened lift delay. Then, when transporting the object to the handover position (transport phase), inaccurate load forces lead to a lower or higher maximum lift velocity. In the subsequent transfer phase, the receiver must produce accurate grip forces in order to achieve a smooth handover and take control of the object completely (Kopnarski et al. 2023b). This means that when lifting objects of unknown weight, both the lift delay and the maximum lift velocity of the person who lifts the object (in handover actions this is the giver) depend on the object weight. Previous studies have shown that people are able to estimate object weight by observing another actor (Hamilton et al. 2007). In a further study, it was shown in a sequential joint replacement task that the second actor shows lower surprise effects in relation to an object weight change than the first actor (Meulenbroek et al. 2007). This indicates that a second actor has information about the object weight from the observation of the first actor. This ability might be used by receivers in handover actions to adjust their initial grip force scaling to the object weight. Assuming so, this would mean that in addition to one’s previous experience (Flanagan et al. 2008, 2009) and cues including size (Cole 2008; Gordon et al. 1991a, b; Li et al. 2009) and material (Buckingham et al. 2009; Flanagan et al. 1995; Flanagan and Wing 1997a), observing other persons’ actions can also directly influence one’s assumptions about an object’s properties and, therefore, their plan during joint actions.

In summary, the knowledge about object weight might be applied by a receiver in the object transfer phase of a handover action. That means that non-apparent object properties, such as object weight, influence the kinematics of an actor during object manipulation (Hermsdörfer et al. 2011; Johansson and Westling 1988; Kopnarski et al. 2023a; Weir et al. 1991). Further, observers seem to be able to determine information about non-apparent object properties from actor kinematics (Hamilton et al. 2007). What is currently unknown is whether this ability to anticipate object properties by just observing an actor is also employed in the observer’s action planning and whether receivers use this in handover actions. This successful anticipation would be evident in adjustments to actions related to object weight. Thus, this would be measurable in the receiver’s peak grip force rate in the transfer phase of the handover action.

The aim of this study was to investigate whether receivers in handover actions can anticipate the weight of an object by observing the giver’s kinematics and use this information to adapt their own action execution in relation to object weight. Hence, we studied whether the giver’s observable action execution (lift delay, maximum lift velocity) differs systematically as object weight is varied. We hypothesized that (1) the lift delay of the giver increases with increasing object weight, and (2) the maximum lift velocity of the giver decreases with increasing object weight. Further, we examined whether object weight is appropriately estimated by the receiver based on the observation of the giver’s movement. For this reason, we investigated the peak grip force rate of the receiver within the object transfer phase as a function of object weight. We hypothesized that (3) the receiver’s peak grip force rate increases with increasing object weight. As a proof of concept, the peak grip force rate of the giver was in addition investigated. Further we expected that (4) the giver’s peak grip force rate is higher for large objects than for small ones and that weight has no effect on giver’s peak grip force rate.