Children’s use of everyday artifacts: Learning the hidden affordance of zipping

Abstract

The everyday world is populated with artifacts that require specific motor actions to use objects as their designers intended. But researchers know little about how children learn to use everyday artifacts. We encouraged forty-four 12- to 60-month-old children to unzip a vinyl pouch during a single 60-s trial. Although unzipping a pouch may seem simple, it is not. Unzipping requires precise role-differentiated bimanual actions—one hand must stabilize the pouch while the other hand applies a pulling force on the tab. Moreover, kinematic data from six adults showed that the tolerance limits for applying the forces are relatively narrow (pulling the tab within 63° of the zipper teeth while stabilizing the pouch within 4 cm of the slider). Children showed an age-related progression for the unzipping action. The youngest children did not display the designed pulling action; children at intermediate ages pulled the tab but applied forces outside the tolerance limits (pulled in the wrong direction, failed to stabilize the pouch in the correct location), and the oldest children successfully implemented the designed action. Findings highlight the perceptual-motor requirements in children’s discovery and implementation of the hidden affordances of everyday artifacts.

1 |. INTRODUCTION

The world is filled with artifacts designed for intended actions and consequences. Think of children’s interactions with toys (interlocking Lego bricks), food (peeling the foil on a yogurt container), toiletry items (pressing the pump of a hand lotion), household objects (twisting a faucet), and clothing (zipping, buttoning, snapping). Such designed actions are so deeply engrained in adults’ activities of daily living that the perceptual-motor requirements seem intuitive. However, for children, it takes years before they learn the specific motor actions to operate everyday artifacts.

Despite the prevalence of artifacts in children’s everyday lives, researchers know little about how children learn designed actions. Previous work focused primarily on age norms for skill onset based on children’s success using zippers, buttons, and so on (Folio & Fewell, 2000; Teaford, 2010), but developmental change in the perceptual-motor requirements remains uncharted. Consequently, researchers cannot know why implementing everyday artifacts takes years to learn; and parents, teachers, and occupational therapists must rely on artistry and common sense to help children perform activities of everyday living. In addition to age-related changes in success, task analyses of particular designed actions and detailed characterizations of age-related changes in the process of discovery and implementation are needed.

Often, the actions required to use everyday artifacts entail only one correct solution, but every artifact affords multiple actions (e.g., grasping, banging, shaking). For a child—or any neophyte who does not already know how to operate the object—discovery of the specific perceptual-motor requirements is elusive because designed actions (pulling open a Tupperware lid) compete with non-designed actions (banging the container). To further complicate learning, typically the designed action is not readily specified by perceptual information (Norman, 1999, 2013). It is “hidden” and users must discover it (Albrechtsen et al., 2001; Gaver, 1991; Hartson, 2003; Rachwani et al., 2020). Discovering the hidden affordances of artifacts can be hampered by lack of perceptual feedback (tight jar lid does not move when twisted, so children cannot discover the designed action of twisting) or by misleading perceptual feedback (lid twists in both directions, impeding discovery that it must be continually twisted to the left).

Moreover, successful implementation often involves complex, motorically challenging operations. Users must know the detailed biomechanical requirements of the designed action and possess the bimanual coordination, dexterity, and strength to implement them (Lockman & Kahrs, 2017). For example, pulling and stabilizing actions are in infants’ repertoires by 12 months of age (Fagard & Lockman, 2005; Kimmerle et al., 2010), but 21-month-olds do not display the bimanual coordination to open the friction-lock lid on a Tupperware container (Rachwani et al., 2020). By 30 months, children learn the specific biomechanical requirements of the designed action—pulling a corner of the lid while stabilizing the base—and only then can they successfully open the container.

Thus, to discover and implement the designed actions of artifacts, children must know how to operate the object and have the perceptual-motor skills to successfully execute the designed action. But paradoxically, without successfully implementing the designed action, children may not discover the specific perceptual-motor requirements. In this study, we conducted a task analysis and investigated children’s efforts to open a zipper pouch—a seemingly simple device patented to be a reliable way of binding the edges of flexible material. Our zipper task was quick with a clear goal: In a single 60-s trial, we encouraged 12- to 60-month-old children to unzip a transparent pouch to retrieve a toy inside. The age range spans infancy, when children display role-differentiated bimanual actions, where each hand performs different but complementary actions (Kimmerle et al., 2010), to school age, when self-care is imperative. Notably, we selected an age range when some children would fail to unzip and some would succeed, so we could observe the underlying behavioral processes. For practical reasons, we used a small pouch instead of clothing worn by the child, and we focused on unzipping rather than zipping because children find it more motivating to open than to close something.

Our task analysis suggests that a zipper serves as an ideal model to examine children’s learning of designed actions. Unzipping requires role-differentiated bimanual actions (Nelson et al., 2013)— one hand must grip and pull the tab while the other stabilizes the fabric of the pouch (Figure 1a–c). But it is not that simple. The perceptual-motor requirements of a zipper are unique to the artifact and are not easy or straightforward to implement. Based on six adults’ success at unzipping the pouch under various conditions (see Supporting Information), we determined tolerance limits for successful implementation (Figure 1a). The gripping hand must apply a force to the tab opposite to the direction of the force exerted by the stabilizing hand, by dragging the tab toward the teeth; if not, the zipper remains zipped. A tab-to-teeth angle >49° decreases the chance of opening, and angles >63° preclude opening (Figure 1a). Moreover, due to flexibility of the pouch material, the stabilizing hand must be below the slider while pulling the tab, specifically within 4 cm of the slider (Figure 1a). If the stabilizing hand is further to the side and/ or bottom, the pouch bends and reduces the probability of opening because the pulling force no longer opposes the force of the stabilizing hand and is not parallel to the direction of the teeth (Figure 1b). If the stabilizing hand is >11 cm from the slider, the pouch bends even more and is impossible to open (Figure 1c). Thus, we examined children’s unsuccessful and successful attempts at pulling and stabilizing.

FIGURE 1.

Specific motor requirements for unzipping a zipper, (a–c) Line drawings of an adult unzipping a zipper pouch. (a) Tolerance for unzipping the pouch based on adult data (see Supporting Information). Colored angular slices indicate whether the pouch will unzip based on the angle of the pull tab relative to the zipper teeth on the pouch. Colored squares indicate whether the pouch will unzip based on the location of the stabilizing hand. (b) Stabilizing hand slightly lateral from the slider during pull of the tab, leading to bending of the pouch. (c) Stabilizing hand at farthest location from the slider, causing the pouch to double over. (d) Developmental and real-time progression of actions with the zipper pouch. Rows represent each child’s timeline for the 60-s trial and rows are ordered by age (rounded to the nearest month). Orange colored boxes behind age numbers represent girls and aqua colored boxes represent boys. Reading down rows shows the developmental progression; reading across rows shows the real-time progression. Timelines <60 s correspond to children who successfully unzipped the pouch. Light gray bars denote durations of time when one or both hands touched either the tab or the pouch; dark gray bars denote durations of time when children held the pouch with one hand and the tab with the other—but do not necessarily denote times when children pulled the tab and stabilized the pouch. Colored lines denote the onset of discrete events when children displayed any pulling action on the zipper tab and any stabilizing action on the pouch. Red lines denote incorrect pulls (tab ≥90° relative to zipper teeth). Yellow lines denote correct pulls (tab <90° relative to zipper teeth) with stabilizing hand in incorrect location. Green lines denote correct pulls with stabilizing hand in correct location. (e) Percent of trial time holding the pouch with one hand and the tab with the other. (f) Total number of pulls (color coded as in timelines). (g) Percent of total number of pulls in correct direction. (h) Percent of correct pulls with stabilizing hand in correct location. Each bar in e–h represents one child.

2 |. METHODS

2.1 |. Participants

We recruited forty-four 12- to 60-month-olds (21 girls) from the New York City area (see y-axis of Figure 1d). We sampled more sparsely at older ages because children immediately succeeded at unzipping. Parents reported children’s race as White (73%), Black (2%), Asian (2%), multi-race (16%), or chose not to respond (7%); 84% were non-Hispanic, 9% were Hispanic, and 7% chose not to respond. Families received a tote bag and photo magnet for participation. Data from three additional children (14–30 months) were excluded due to fussiness (n = 1) or experimenter error (n = 2). All children were healthy and born at term. Parents reported whether children had unzipped a zipper in the past week. However, children’s experience did not correlate with their age or success at unzipping, ps ≥ .202, so we concluded that parents’ reports were unreliable.

2.2 |. Procedure

The zipper task was administered between trial blocks of other studies on manual actions (using a hammer, opening cabinet latches, and so on). Children sat at a child-sized table or in a highchair across from the experimenter. Caregivers sat behind children, filling out a questionnaire.

The zipper pouch was rectangular (12.5 × 5.5 cm), made from vinyl, transparent, and had a small toy inside to motivate children to open it. The tab was enlarged (2 × 5.5 cm) and colored red to ensure its saliency. Trials began when the experimenter placed the pouch on the table, within arms’ reach of the child. However, 8/44 children quickly grabbed the pouch before the experimenter could place it on the table. For 27 children (M = 31.4 months), the experimenter offered the pouch with the tab closer to the child’s left hand, and for 17 children (M = 31.8 months), with the tab closer to the child’s right hand. At the beginning of the trial, the experimenter asked, “What can you do with that?” The experimenter then looked down to fill out “paperwork” to justify a break in social interaction. If children sought help, the experimenter encouraged them to continue, “This is for you to play with.” If children bid to caregivers, the experimenter reminded caregivers not to provide instructions. The trial ended after 60 -s or when children opened the pouch.

Two fixed cameras recorded children’s hands and face from front and side views. Two additional fixed cameras recorded the experimenter and caregiver. All four views were mixed into a single video frame for ease of coding.

2.3 |. Data coding

Coders scored videos frame by frame using Datavyu software (www.datavyu.org). They identified each time children held either the tab, the pouch, or both. When children held both the tab and pouch, coders determined if children pulled the tab, and whether the pull direction was correct (pulled tab toward the teeth, i.e., <90° relative to the teeth) or incorrect (pulled tab away from the teeth, i.e., ≥90° relative to the teeth). For every pull, coders scored the location of the stabilizing hand (in a 4-cm wide column directly below the slider-teeth or any other location).

Different coders rescored 100% of data to ensure inter-observer reliability. Coders were in exact agreement on ≥95.3% of frames (κs ≥ 0.92, ps < .001) for holding the tab and/or pouch. Coders agreed on ≥86.7% (κs ≥ 0.83, ps < .001) of times when pulls occurred, pull direction, and location of the stabilizing hand. Disagreements were resolved through discussion.

3 |. RESULTS

As shown by the timelines in Figure 1d, we analyzed each child’s data in real time to characterize the process of discovery and implementation and to determine the causes of children’s success or failure. Moreover, we tested change in each task component by age (Figure 1e–h). We used Spearman’s rank correlations because not all task component variables were normally distributed (based on Shapiro-Wilk tests). Preliminary analyses showed no effects of sex or starting pouch orientation across children, ps ≥ .132, so these variables were collapsed for subsequent analyses.

3.1 |. Success and time to unzip

Children appeared highly motivated to open the pouch to retrieve the toy. However, 19/44 children did not succeed at unzipping—not even the slightest bit. All 11 children <24 months failed to unzip (shown by 60-s bars in Figure 1d). Between 24 and 34 months, 7/16 failed, but at 35+ months, only 1/17 children failed. Despite failure to retrieve the toy, children were highly engaged to perform non-designed actions—rotating and manipulating the pouch, wiggling the tab, and so on (see times children touched tab or pouch in Figure 1d).

Of the children who successfully opened the pouch, time to unzip averaged 13.4 s. Between 24 and 34 months, four successful children unzipped in 6.8–10.06 s, and five unzipped in 15.3–56.5 s. Children ≥35 months unzipped in 2.2–8.3 s, with the exception of one 37-month-old (15.4 s) and one 48-month-old (23.1 s). As shown by the length of their timelines in Figure 1d, time to unzip decreased with age: With all children included, r_s(42) = −.78, p < .001; with only successful children included, r_s(23) = −.67, p < .001.

3.2 |. Holding pouch with one hand and tab with the other

Light gray bars in Figure 1d show that children sometimes held only the pouch with one or both hands (M = 38.6% of trial time), but rarely only the tab (M = 3.5% of trial time). With the exception of one 14-month-old, at some point in the trial, all children held the pouch with one hand and the tab with the other (dark gray bars in Figure 1d–e). Children displayed pouch-tab holding during M = 45.0% of the trial, and percent of time with pouch-tab holding increased from 0%–38% in children <24 months, to 22%–69% in children between 24 and 34 months, to 33%–84% in children ≥35 months, r_s(42) = .53, p < .001.

Every switch from one dark gray bar to the next in Figure 1d indicates a change in pouch-tab holding—children turned or flipped the pouch to access a different vantage point, or swapped hand roles so the tab hand became the pouch hand and vice-versa. Overall, children displayed M = 5.64 bouts of pouch-tab holding (range = 0–20 bouts), and frequency of bouts decreased with age, r_s(42) = −.37, p = .012. That is, older children reconfigured their hands fewer times.

3.3 |. Pulling the tab

Despite pouch-tab holding, children did not necessarily display the required pulling and stabilizing actions. That is, children’s hands sometimes played distinct roles (pulling tab and stabilizing pouch) and sometimes not (merely holding both parts). Vertical lines within the dark gray bars in Figure 1d and the histogram in Figure 1f show pulling actions on the zipper tab (M = 5). With the exception of the three youngest children (12- to 13-month-olds) and one 19-month-old, all children pulled the tab at least once. Children <24 months displayed 1–8 pulls, children between 24 and 34 months displayed 1–24 pulls (7/16 pulled more than 10 times). Like the youngest children, those ≥35 months displayed 1–8 pulls. As displayed in Figure 1f, number of pulls did not increase with age, r_s(42) = −.06, p = .678, most likely due to fewer pulls in the youngest and oldest children compared to children in the middle age range.

3.4 |. Pulling in the correct direction

Despite multiple pulls, children frequently pulled in incorrect directions away from the teeth. Green and yellow lines in Figure 1d and the histogram in Figure 1g show times children pulled in the correct direction—<90° relative to teeth—whereas red lines denote times children displayed incorrect pulls (≥90° relative to the teeth).

Children <24 months rarely pulled correctly (0%–13% of pulls), children between 24 and 34 months displayed more correct pulls (0%–100% of pulls) and half of them (8/16) pulled correctly >50% of times. Children ≥35 months pulled correctly 2%–100% of pulls and 11/17 of them did so on 100% of pulls. Thus, children showed a developmental shift from pulling in incorrect directions (M = 66.9% of pulls), to pulling in the correct direction (M = 68.5%); reading down the rows in Figure 1d shows an age-related shift from red to yellow and green lines. The number of correct pulls relative to all pulls increased with age, r_s(31) = .54, p = .001. The lack of a consistent color switch from red to yellow/green in individual timelines indicates that children did not improve from pulling in incorrect to correct directions in real time.

3.5 |. Stabilizing hand location

Even when children pulled in the correct direction, they sometimes placed their stabilizing hand too far from the slider, causing the pouch to bend. Green lines in Figure 1d and the histogram in Figure 1h indicate times children appropriately placed their stabilizing hand when pulling correctly (M = 78.3% of correct pulls). Yellow lines indicate inappropriate placement of the stabilizing hand (M = 21.6% of correct pulls). When pulling correctly, children <24 months never placed their stabilizing hand correctly; children between 24 and 34 months placed their stabilizing hand correctly 0%–100% of times and 8/16 did so >50% of times; with the exception of one 48-month-old, all children ≥35 months, placed their stabilizing hand correctly. Thus, scanning down the timelines in Figure 1d and shown in Figure 1h, correct hand placement relative to correct pulls increased with age, r_s(31) = .62, p < .001. Intermingling of green lines with red and yellow within individual timelines indicates that children did not learn the precise perceptual-motor requirements within the trial, and repeated green lines within individual timelines indicates that children did not exert sufficient force to move the zipper tab along the teeth.

4 |. DISCUSSION

The designed actions of everyday artifacts are prime examples of hidden affordances—the goal is clear, but the biomechanical requirements to implement it are not. We encouraged children to unzip a pouch to examine developmental change in the perceptual-motor skills involved in discovery and implementation of the designed actions of everyday artifacts. Unzipping the pouch required pulling and stabilizing simultaneously with the hands in the correct locations to exert forces in opposite directions.

Pulling and stabilizing are common actions in children’s everyday activities. Children display one-handed pulling actions by 6 months (CPSC, 2020). They display consistent role-differentiated bimanual actions, including stabilizing an object with one hand while operating on it with the other hand by 13 months (Kimmerle et al., 2010). Nonetheless, despite our efforts to make the task easier (e.g., enlarging the zipper tab to remove the need for a pincer grip, using a small pouch that children could hold and stabilize in one hand), most children did not successfully coordinate pulling and stabilizing actions with our zipper pouch until 24+ months of age. Unzipping their own coat occurs later, between 32 and 40 months (Teaford, 2010). And although children may have had prior experience unzipping, they still had to learn the specifics of implementation for our zipper pouch. Indeed, it took children several attempts to unzip our pouch, likely because they had to discover the unique requirements of the specific zipper. Presumably, children learn the perceptual-motor requirements for all designed actions on a similar object-by-object basis (Rachwani et al., 2020). Children may learn that a zipper requires pulling and stabilizing, but it takes many more months for children to recognize that the zipper on a plastic bag, purse, or coat requires subtle adjustments of the same actions, and to learn to make the necessary adjustments in real time as they encounter new instances of similar artifacts.

As in previous work with twist-off and pull-off lids (Rachwani et al., 2020), the real-time process of unzipping undergoes significant, protracted developmental change, from non-designed actions, to display of the designed action, to successful implementation. The youngest children did not even know to pull the zipper tab. Pulling was infrequent and children’s hands were not on the tab and the pouch. Instead, they looked at the toy, manipulated the pouch, wiggled the tab, and so on. Children at intermediate ages knew to unzip, and thus pulling was frequent. However, many struggled with the biomechanical requirements by pulling in incorrect directions and/or stabilizing the pouch in incorrect locations. Pulling the tab toward the zipper teeth is not intuitive, and pulling the tab away from the teeth is likely more perceptually salient, as if pulling the tab and pouch apart. The oldest children quickly placed their hands appropriately—the gripping hand pulled the tab by applying a force toward the direction of the teeth, and the stabilizing hand applied an opposite force below the slider.

Thus, the designed actions for many artifacts, including toys designed for young children, are motorically difficult to discover and implement and take years to achieve (Jung et al., 2018; Ornkloo & von Hofsten, 2007; Ossmy et al., 2020; Rachwani et al., 2020). However, despite similarity of outcomes among children of similar ages, our single zipper trial might not represent children’s true potential. Moreover, many designed actions do not allow for one-shot learning. Children showed no evidence of learning in real-time. They did not go from pulling in incorrect directions to pulling in the correct direction within the trial, although success may be underestimated because of the arbitrary 60- s trial length. In addition, given that children received a single trial with the pouch oriented one way only, outcome measures might have been biased by pouch orientation. However, children’s spontaneous interactions with objects in everyday life are likely isolated and brief because of the immense number of options in their surroundings.

Precisely which type of factors facilitate discovery and implementation of the perceptual-motor requirements of everyday artifacts remains to be examined. Whether prior experience, verbal guidance, or physical assistance contribute to learning the perceptual-motor requirements is still unknown. Regardless, children cannot benefit from the artifacts that populate their everyday world without knowing the designed action and possessing the perceptual-motor skills to implement it. Both are slow to develop because lack of motor skill impedes discovering the designed action and knowing what to do does not guarantee successful implementation.