This paper was presented at the Notation in Contemporary Music Symposium at Goldsmiths College on 19 October 2013.
Notation varies considerably in the way it prescribes actions and results. While some approaches tend towards precision and the specification of events with great detail, others present generalised fields of possibility. In his ‘A Summary of the Characteristics of Scores’, landscape architect Lawrence Halprin suggests that ‘Scores themselves open up options rather than closing them down’. While this seems on the face of it to be counter-intuitive – scores by their very nature define, or fix things – they cannot be exhaustive in their directions to people choosing to make realisations of them. While many scores propose specific actions, they might also define a space, or process, or set of circumstances. Different approaches tend more or less towards specificity or generality.
It is the latter that currently interests me as a composer. I find myself always trying to find the most general way of framing a piece at the moment. When planning a piece, there are so many ways in which the basic concept could be articulated, normally through specifying certain parameters, durations, actions, or sounds. When considering these possibilities, I am always conscious of their arbitrariness. I could do one thing, or another, or something entirely different, and the piece would still represent the idea I wanted to work with. This is something that has been central to my interest in open forms for some time now. In previous pieces, I have tended to leave various things open, such as instrumentation, selection of material from a field of possibilities, use of time windows to reduce the amount of synchronisation between sounds, and distributed decision making. The purpose is to make available more of the possible ways a piece may be constituted. Why is a single, fixed version better than any other? For me it’s not. It is simply one possibility amongst many. Of course it’s possible to generate a justification for a particular choice, and composers do this in various ways, such as framing decision-making as a personal choice, or through a process, or more complex rationales. Ultimately they are self-determining; I invent a system to make this piece, and the piece works because it embodies the system. I am right because I made it, and that’s absolutely fine.
So recently I’ve been more aware of the need to consider the point at which I intervene to make the piece. I am trying to find the most efficient way of presenting the material to allow the idea to be realised, but without including any arbitrary choices. This seems to involve removing information about sounds in particular, and on occasion ways of dealing with time. On the whole this seems to work for the pieces I am making at the moment, where a process or activity is at the heart of the work, or where there is something specified which shapes the production of sound such that anything within the range of possible sounds that could be produced is acceptable. I wonder if this enough though. Is acceptable really sufficient? At the back of my mind I still have this notion that what I do as a composer should involve making scores and specifying sounds with respect to time. My job is, supposedly, to show masterful control of my material to produce novel sonic effects. I guess I could try to do that, but it comes down to specifics again. The quality of a sound or its context can be controlled in fairly specific ways, whilst retaining a wide degree of choice. All sustained sounds lasting over thirty seconds have something in common, as do all extremely high sounds, or those on the edge of audibility. I think it’s these generic characteristics which are most useful for me as parameters now as it removes the minutiae and irrelevance of decision-making. The general provides specificity, more so than the assumptions inherent in the precise.
In this paper I will explore ways of considering notation from a generalised perspective, and how such an approach might question the nature of arbitrariness in compositional decision-making. I will look at two examples of this in work by others, and how taking a generalist approach worked in one of my own pieces.
The questioning of specifics and the move towards generality can be seen in the development of Fluxus artist George Brecht’s pieces Time-Table Music (1959) and Time-Table Event (1961). Time-Table Music (1959) is among the first group of Brecht’s event scores, composed during his participation in Cage’s composition class at the New School for Social Research between 1958-59. Like some of his other pieces composed at this time, including Candle-Piece for Radios, Card Piece for Voice (both 1959), and Spanish Card Piece for Objects (1960), it uses found materials to structure a performance:
George Brecht: Time-table Music (1959)
Here the use of timetables provides a chance-determined time structure as their original purpose and function has been replaced by a newly contextualised role. There is though still an implicit shape that might emerge based on typical journey schedules: there are likely to be fewer journeys in the very early morning for instance, with realisations of the piece tending more towards medium durations, and capping them at 23’59”. So Brecht’s choice of materials shapes the realisation, something that he observes in his essay Chance-Imagery (1957):
It is sometimes possible to specify only the universe of possible characteristics which a chance event may have. For example, a toss of a normal die will be expected to give a number from one to six. Any particular face will be expected to turn up in about one-sixth of a great many throws. But the outcome of any one toss remains unknown until the throw has been made. It is often useful to keep in mind this “universe of possible results,” even when that universe is hypothetical, for this clarifies for us the nature of our chance event as a selection from a limited universe.
Early drafts of the piece show how both the language and its implied performance practice developed during its composition, gradually becoming more precise. The first sketch of the score, dated 13 July 1959, contains the main components of the piece, that of obtaining timetables and using them to structure a period of time and control sound production by the performers. At this stage though, the location of the performance is open. Brecht suggests that
The piece may be performed in:
a classroom, [on?] pencils, pins, erasers, etc.
a restaurant, [on?] knives + forks, etc.
This draws on the immediate everyday environment. The possibility of various performance locations was under consideration at this stage, with Brecht indicating on the revision the following day that it was ‘To be performed in a railway station.’ and then in the following paragraph ‘The performers enter a (public place)’. The range of possibilities was clearly discussed with other class members, as on both these sketches he indicates that ‘For a later generalization of this piece. (I am indebted to A.[lfred] Hansen for an element of this piece: the nature of the many places in which it may happen.)’ . The following two pages of the notebook seem to have resolved this question however. In a handwritten version of the score on page 107, dated 24 July 1959, Brecht now states that ‘The performers enter a railway station…’, and this is confirmed by a typed copy of the score stuck to page 106, and Time-Table Music was first performed by members of the class at Grand Central Station in New York. This move from the general to the specific, thereby closing off potential realisations, is perhaps at odds with Brecht’s later work where the ambiguity of single words or groups of words opens up the realisation to more possibilities.
Elsewhere in the drafts of the score, ideas about the performance practice can be found. In an undated entry on page 106, Brecht suggests some ‘Ideas arising in connection with TT Music’, specifically questioning ‘How “together” are the performers in space? (implies: who are the listeners?) (i.e. do the performers wander all over the station making sounds?’ and then ‘Comment: Ambiguities: 1. Are the performers free to move about the station once they have begun together?’ So he is aware that there are ambiguities present, and that these create an interpretative space. Crucially, none of these thoughts find a place in the final score. They are not ultimately central to the essence of the piece: through the published version, it remains possible for the performers to move, but they need not. This is an example of the score as a focus, not as a means of controlling it completely. The piece must however be performed by a group and the emphasis on ensemble performance is clear in the instruction ‘They stand or seat themselves so as to be visible to each other, and, when ready, start their stopwatches simultaneously.’ This concerted action is a shared experience (at least initially), and inherently musical.
It should also be noted that even though he refers to listeners in his notes, the extent to which the results of Time-Table Music are perceivable is left unspecified. Although performers are instructed to make sounds, there is no obligation for these to be projected in such a way that the piece becomes noticeable to others. Performances may be potentially extrovert and disruptive, or private and subversive. This range of possibilities is even more open in Time-Table Event, the more generalised, separate piece Brecht developed from Time-Table Music in early 1961:
George Brecht: Time-table Event (1961)
The main difference in the later piece is the move from sound-based music to a duration-determined event. Specifically, Brecht removes any reference to sound. By this time he was ‘…increasingly dissatisfied with the purely aural qualities of a situation … my thought being that the word “event” was closer to the multi-sensory (total) experience I was interested in than any other.’ As with his Fluxus colleagues, the consideration of performance as an intermedial experience was central:
Composers, performers and auditors of music permit sound-experiences by arranging situations having sound as an aspect. But the theatre is well lit. I have to cough; the seat creaks, and I can feel the vibration. Since here is no distraction, why choose sound as the common agent?
So Time-Table Event simply uses an obtained timetable to determine the duration of an event in a railway station. In the later piece the obligation to make sounds, or indeed structure the performance, based on the rows and columns of the chosen timetable is also removed: the score states only that the timetable ‘determines the duration of the event’. This is as a result a very different piece, but one which might subsume the earlier one. It is more generalised.
The performance situation is also reconsidered. Time-table Music states that it is ‘for performance in a railway station’, whilst Event is ‘to occur in a railway station’, although this change only occurred between the two drafts of the later piece. This crucial difference is a result of Brecht’s move towards events as a structure, and away from the need to present work with any kind of overt theatricality. In ‘The Origin of “Events”’ (1970), Brecht states that
In the Spring of 1960, standing in the woods in East Brunswick, New Jersey, where I lived at the time, waiting for my wife to come from the house, standing behind my English Ford station wagon, the motor running and the left-turn signal blinking, it occurred to me that a truly “event” piece could be drawn from the situation.
Events can be observed, or thought about, as well as realised: a jotting on page 20 of his notebook in late March 1961 remarks ‘perceiving is an act’.
The sketches also make reference to the requirement that the tabulated times are ‘chosen by chance’, which is also absent from the final score. Brecht’s extensive research into chance as a formative process for making art is well documented, but the removal of this instruction from the score is as a result of its redundancy: timetables and their use in this piece, to an extent, promote chance inherently. Time-table Event also removes references to the use of stopwatches and any sense of co-ordination between players. This is in part because it can be realised as a solo: there is no indication that it is for a group, as is the case with Time-table Music (the performers, they stand, each performer).
So in this example, Brecht’s approach develops from an idea that remains at the core of both pieces: going to a railway station and using a timetable to determine the duration of an event. In the initial sketches for Time-table Music he appears to be attempting to define a set of specific circumstances to frame the piece, albeit with a number of variables. The reference to Al Hansen’s suggestion the piece could be performed in ‘many places’ suggests that Brecht considered how location might be left open and that the general had a place, but this aspect was later rejected in both the final versions. The role of performance, sound-making, and the detailed structure of events with regards to lines in the timetable all disappear between the two versions however. Their removal might be seen as a result of their superfluity. They were perhaps not central to the core idea for the piece. Although the second score is more open, it is less arbitrary in its construction. The idea is more clearly apparent, and not obscured by as much subjective detail.
The removal of the arbitrary can also be seen in Christian Wolff’s verbal scores. Wolff suggests that determining the identity of a piece is of importance to him, providing only sufficient guidance–and removing unnecessary detail–for that to be made apparent. Discussing the role of indeterminacy in his work, Wolff notes
What I do is think of the worst case given the indeterminate conditions and the freedom which I give to the performers: what could somebody do given the restrictions I’ve set? What’s the worst that they could do from my point of view? If I can accept that, if that’s still okay, then the thing is all right.
Although this comment relates to the degree of openness afforded to performers in his pieces, it impacts on the level of intervention by the composer when framing a piece. This test, when applied by different composers, has the potential to say a lot about their views on generality. For any given piece, we could conceive of a range of possible interventions from the locked and precise specification of micro-events to the openness of broad statements of intent. Taking the example of using stones as a sonic resource as a point of reference for a piece, a number of different approaches have been taken by composers. The Frog Peak Rock Music Book contains 33 pieces which use rocks, stones and pebbles in some way. Although for many of these the use of stones is not essential or even the core idea of the piece, others demonstrate the way in which different levels of specification are employed. A comparison between Wolff’s Stones and Pauline Oliveros’s Pebble Music shows the way in which the two composers choose to intervene in different ways to produce a piece which focuses on making sounds with stones.
Pauline Oliveros: Pebble Music (1992), instructions
In Pebble Music, Oliveros provides a series of requirements for sourcing pebbles, noting fairly precise dimensions, shapes, and densities. She follows this with a list of ‘performance techniques’ for making sounds with the pebbles, and methods for modifying these sounds through playing with resonance. In a performance, this set of techniques is linked to a grid of time periods that specifies how and when to make sounds with the pebbles, and how this should relate to instrumental sound.
Christian Wolff: Stones (1969)
Compare this with Wolff’s Stones, which also suggests there should be a variety of sizes and kinds sourced. Wolff also notes approaches to sound production using stones and other materials, but does not use the same level of specificity as Oliveros. There is also no indication of duration or temporal structure to explain how these sounds might relate to each other.
Neither score is better or worse, but they say something about the priorities of each composer. Both pieces involve experimenting with the sonic properties of stones in relation to other materials. Oliveros focuses on resonance and instrumental doubling, in addition to specifying five types of sound production. This perhaps implies these are the only five methods permissible in the piece, and that rubbing, clicking, shaking, dropping and rolling, in combination with different types of resonance and instrumental imitation, define the piece in some way. In the Wolff, there are some limitations, such as making mostly ‘discrete sounds’ that are ‘sometimes in rapid sequences’. He also specifies modes of sound production, but does so in a less prescriptive manner, permitting ‘striking stones with stones’, ‘stones on other surfaces’ and ‘other than struck’. The brevity of the score and lack of further definition of how these actions might be made is in contrast to Oliveros, who includes information about varying regularity, distance, pitch, resonance, and amplitude.
Aside from the principal differences between the two pieces–the use of musical instruments alongside the stones–they are very similar. Oliveros guides and limits the instructions for exploring the sound of stones more than Wolff, who relies on the performer to make these explorations within broader categories. Oliveros presents a time structure for deploying these sounds, and Wolff does not. Both approaches create different identities and emphases, with Wolff noting that Stones is ‘an extreme instance of combining maximum transparency, flexibility and freedom for the performers with at the same time an unmistakable, irreducible identity.’ 
The projection of ‘identity’ is central to how decisions about notation and instruction are made. While the ‘Wolff test’ could be thought of as a tool for composers to consider what is right in a piece, the definition of ‘right’ is more difficult to articulate. Wolff is clearly concerned with specificity and variability. Elsewhere he observes
One could also ask, what is the specificity that is notated. The notation tells you play c, f sharp and b at dynamic level ‘pp’. That’s clear and specific enough, but the musical sense? I tell you ‘make sounds with stones’. No pitch, duration, rhythmic procedures, dynamics. But the piece is quite specific about the sound itself, the stones – that will be its identity. To be sure considerable stretching is possible. Usually when the piece is played, though its overall durations is not specified, it will run from 5 to 10 minutes. There is though a recorded performance of it that runs for an hour, including (because of an agreement among the performers that each would make no more than 15 sounds over the total time) vast spaces of silence. Very different versions of the same score, but always the specific identity of the performed sound of stones (only). This kind of variability has long interested me, and the verbal scores are a way of realizing it.
In some of my recent pieces, considering generality has been a pressing concern. I have noticed that the composition process, for me, tends to start with a relatively clear idea about what will happen in a piece. There is then a period of work where I consider how to articulate this idea, and then the score is made. It is in the central period where I’ve noticed my practice has changed considerably. There are still vestiges of my previous working methods, generally involving the arbitrary construction of processes to determine the characteristics of events, but my focus has been on stripping these away where they are not relevant to the original idea, where they get in the way.
For example, in things whole and not whole, a piece for a large group of musicians written for the Basel Sinfonietta in 2011, the final notated version of the score went through a number of changes. These were driven by the different groups who subsequently performed the piece, as well as my desire to clarify the core idea. My starting point for the piece was to create a cueing network modelled on research into the behaviour of flocking birds carried out by Craig Reynolds in the mid-1980s, which resulted in a computer simulation of flocking boids. The movement of Reynolds’ boids was governed by three simple rules: a clumping force which keeps the flock together, a separation force which keeps them apart, and an ability to match velocity. The individual boids follow these rules, and the flock movement emerges from their interaction.
In things whole and not whole, the interaction of the players is governed by rule-based cueing: players make sounds as soon as possible after those made by other players, choosing different reference players for each new sound they make. The core idea of the piece was determined fairly early on. I mapped Reynolds’ rules to equivalents that might govern ensemble interaction with sound over time.
James Saunders: things whole and not whole (2011), first version
Clearly these sounds needed to be articulated in some way. This was a commission for an orchestra I hadn’t worked with before and knew very little about in terms of their likely acceptance of alternative approaches to scoring. My initial work on the piece was largely governed by this perceived constraint therefore. How could I get the idea across such that it sounded as I intended, but without alienating the players unnecessarily?
James Saunders: things whole and not whole (2011), initial score mockup (with previous title)
The first drafts of the piece involved creating categories of instrumental sound, creating groups of pitches across the entire range of each instrument using ‘normal’ and unconventional sound production, split between the constituent members of each instrumental group. Selections from each pitch/timbre set were distributed across the individual parts in one of three durational categories: ‘as short as possible’, ‘long (c.10”)’ and ‘very long (c.30”)’. I was very much aware of the imbalance of sound types across a standard symphony orchestra, with over half of the sounds being made by string instruments. I wanted to find a way to homogenize the palette without reducing the variety: it needed a more even spread of timbres. Initially I explored expanding the instrumental resource by allocating objects to each player alongside their regular instruments as I had done in a previous orchestra piece, geometria situs (2010), which went some way towards reducing the prominence of the string timbres.
I felt it needed to go beyond this however. I discussed how a more open consideration of the sonic palette might be received by the orchestra through an email correspondence with one of its members, whom I had worked with before and who understood the dilemma. This encouraged me, so I removed all the conventional orchestral instruments and began making a series of actions involving objects and materials. I removed all the pitched material and kept only the specification of objects, but began to do this in a more specific manner. I created a list of over 200 actions, such as ‘crumple cellophane into ball’, ‘pour salt onto glass’, and ‘operate electric drill’, but realised the futility in trying to map out what was a vast range of possibilities. I was concerned about the arbitrary nature of my decision-making. Throughout both attempts, selecting pitches, playing techniques, and objects was somewhat unsystematic. I considered creating systems to make these decisions, but they too would be arbitrary. I needed a way to maximize the variety of sound selection without it being constrained by my own particular preferences and what I had to hand.
As with an earlier piece, imperfections on the surface are occasionally apparent, distributed choice was the answer. There is a lot of potential in asking the 60 orchestral musicians to each select their sounds independently, so the final version for Basel asked them to do just that. Each player had to find 16 sounds, each of which should, as far as is possible, be different to those used by the other players. This resulted in 960 ‘different’ sounds. I was aware that this might be too open for some of the players, so I additionally gave them some directions to help select materials and guidelines and rules to remove certain possibilities. I specified ‘noise sounds’ and asked the players to ’emphasize and prioritize noise components over pitch in each sound’, primarily to avoid focus on pitch patterning, which would have distorted the idea I had for the piece. I gave the performers the option to use any combination of their normal instruments and found materials, recognizing that some players might be resistant to abandoning something that was central to their identity as a musician.
On completing this sourcing of materials, I asked the players to notate their 16 actions in an appropriate way using a blank score page template, and also provided an example. This material was then used to articulate the piece’s core rules:
James Saunders: things whole and not whole (2011), first version player template
I attended the first rehearsal in Basel and was both surprised and pleased to see that virtually everyone had abandoned their instruments, and the stage was strewn with household debris and the occasional double bass or trombone. The decision by these players to use their instruments now felt like a conscious choice. There was a lot of variety: balloons, umbrellas, bottles, food, kitchen implements, sports equipment and building materials were abundant. In a break, one of the string players approached me and asked if it was acceptable to use a lump of uranium he had extracted from an open mine in Germany, and a Geiger counter. Suffice to say I would not have envisaged this as a possibility myself.
The notation ‘worked’ in that it was functional and encouraged a group of orchestral musicians to experiment with found materials to make sound. Shortly after this performance, another opportunity arose for the piece to be played in London by a group of about 30 musicians who specialise in experimental music and improvisation. I looked at the score again, and decided that much of it was superfluous for this different context. I decided to remove the score page as it was unnecessary; in fact, very few of the Basel players seemed to use it in any case, and it became apparent that most were using more than 16 sounds. So the new version of the score comprised only the preparation and instructions texts, as well as the page listing the suggested sound types, which now admitted electronic sources. A comparison between the two ‘preparation’ texts show some other small changes.
Although the piece is largely the same, the score for the second performance removes much of the superfluous information and further generalises it. It worked perfectly well in performance, but afterwards I looked through what I presented to the musicians once again with a view to seeing if it could be refined further. It still felt somewhat unwieldy and overwritten as a set of instructions, and I wanted to tighten it up to avoid including information which was either unnecessary or in itself created ambiguity. For example, by providing a list of suggested sound sources, it implied preference for these approaches. The final version of the score is therefore much more concise:
James Saunders: things whole and not whole (2011), final version
The piece is now perhaps less explicit in some of its instructions, especially in relation to performance practice where implication and a tacit understanding of the idiom is taken for granted; examples include no mention of visual cueing and preparing sounds in advance. Other changes are more significant, such as the removal of the requirement to match volumes which proved to be impractical and of insignificant value. The piece now exists in a more generalised form, and its proximity to my initial sketches for the piece is readily apparent. I removed a lot of the detail in order for the premise to be clearer.
It follows though that by removing information something is lost. You could imagine a process whereby any piece is eroded to the point where nothing remains except a rather generic instruction for some kind of activity (although this is very successfully taken to an extreme in some of Brecht’s scores). Clearly I’m not proposing that this is necessarily always a good thing, that all pieces have too much information or character, or that detail and specificity somehow get in the way for all music. For some pieces, by some composers, considering the most reduced version might be a useful strategy. But what was lost in this process for me? In the example of things whole and not whole I would argue not a great deal, while at the same time clarifying what the piece is and communicating that in a more direct manner. The point of clarification was my realisation that I didn’t need to specify all the variables and control that at the micro level, rather that through opening those decisions up within a large group of people I could create the variety I required in a more engaged way. I replaced a large number of arbitrary decisions by me, including pitch systems and personal sound choices, with a clear principle and method which produced arguably more variety and rigour. It might be argued that in doing so I ceded some of my responsibility as a composer, and with it the detail and ‘character’ of the piece formed by my own idiosyncrasies. While this may be the case, I feel the solution replaces this with a more generalised approach, that of distributed decision-making. It puts people and massed individuality at the heart of the work.
Of course such an approach works best where concepts, processes and strategies are central to a piece’s operation. It is implicitly more suited to verbal notation than to stave or tablature for instance, given the latter’s inherent focus on the precise. But this also questions the appropriateness of certain notational strategies and our reasons for selecting them. The gradual editing out of stave notation, via focused instructions, to more general principles expressed through words allowed me to communicate only what I felt was necessary for the piece. Stave notation proved a distraction as my ideas for the piece clarified. I was trying to notate a specific version of the piece, albeit with many variables, when the general situation was sufficient. It passed the Wolff test too, I hope, on each occasion, with its identity being clear despite the local differences created through individual choice.
On reflection, I wonder if this is not in fact about generality, but rather as Wolff suggests to do with creating identity for a piece. It’s easy to get caught up in decision making, invention and designing processes and systems without any regard for the underlying principle. We make so many arbitrary decisions as composers, and it’s the reason we make such a wide range of music, but sometimes these can be subsumed by taking a more generalist position to allow identity to be projected with greater clarity and, on occasion, character.
 Lawrence Halprin, ‘A Summary of the Characteristics of Scores’, The RSVP Cycles: Creative Processes in the Human Environment (New York: George Braziller, 1969), p. 190.
 George Brecht, Chance-Imagery (New York: Something Else Press, 1966), 5
 Brecht, George, Notebook III: April 1959 – August 1959, ed. Hermann Braun (Köln: Walther König, 1991), p. 99.
 Ibid., p.105.
 Ibid., p.105.
 Ibid., p.107.
 Julia Robinson, ‘From Abstraction to Model: George Brecht’s Events and the Conceptual Turn in Art of the 1960s’, October, no. 127 (Winter 2009): 85.
 Brecht, George, Notebook III: April 1959 – August 1959, ed. Hermann Braun (Köln: Walther König, 1991), p. 106.
 It should be noted that both pieces are found in Water Yam, and as such have separate identities.
 George Brecht, ‘The Origin of “Events”’, reprinted in Julia Robinson, George Brecht: A Heterospective (Köln: Walther König, 2006), p. 236.
 Brecht, George, Notebook VII: June 1961 – September 1962, ed. Hermann Braun (Köln: Walther König, 2005), p. 115.
 See Brecht, George, Notebook VI: March 1961 – June 1961, ed. Hermann Braun (Köln: Walther König, 2005), p. 21 for both versions of Time-Table Event, dated 26 March 1961.
 George Brecht, ‘The Origin of “Events”’, reprinted in Julia Robinson, George Brecht: A Heterospective (Köln: Walther König, 2006), p. 236.
 Brecht, George, Notebook VI: March 1961 – June 1961, ed. Hermann Braun (Köln: Walther König, 2005), p. 20.
 Ibid., p. 21
 See George Brecht, Chance-Imagery, (New York: Something Else Press, 1966)
 Wolff, C., ‘I can’t shake Webern’s influence’ in Wolff, 1998, p. 170.
 Wolff, C., ‘Stones’ in Wolff, 1998, p. 494.
 Lely & Saunders, p. 412.
 geometria situs was commissioned by Südwestrundfunk and first performed by the SWR Sinfonieorchester at the Donaueschinger Musiktage on 17 October 2010. All the orchestral players need at least one auxiliary instrument that they must source independently. These are chosen from five specified categories: blown tube; bowed plastic cup; bowed polystyrene; bowed wood; and coffee cup on surface.