These sentences, quoted from the author's Preface, sum up the essence of John Irving's second book on Mozart's piano sonatas. The approach is challenging, and the result yields refreshing thoughts.

The book consists of two parts of unequal length: the longer one deals with the visual study of the score (part 1: ‘Reading Texts’), the shorter one with the actual playing of the music (part 2: ‘Playing Texts’). The second half of part 1 and part 2 as a whole complement each other nicely, as two sides of the same coin. On the one hand, sensible analytical observations and comparisons of autographs and printed editions show remarkable insight into Mozart's workshop; on the other, the ideas and experiences of the practical musician evoke the living organism of the music.

Tackling the book in retrogade order, chapter 7 in part 2, ‘Embellishing Mozart's Texts’, offers original thoughts on the subject. ‘Should We Do It?’ sounds the initial question, in response to which the author fleshes out the details of his argument with a careful sense of responsibility. Important, sometimes disputable, issues arise, and tasteful examples are suggested for embellishments in k457/iii (successive returns of the opening theme), k570/i and k309/ii. Irving's foremost hero in this respect is Robert Levin, of course, but due attention is also given to the recordings of Ronald Brautigam, Tom Beghin and Andreas Staier.

The minutely detailed analysis of the Andante un poco adagio of k309 (126–131) raises two interesting issues. The first, setting out from the well-known pedagogical background of the sonata – which, according to Irving, explains the painstakingly notated variants and embellishments of the movement – proposes that what was prescribed for a teenaged pupil of Mozart might surely be transformed by an educated player: ‘[Mozart's] notated embellishments may be regarded as exemplary … ways in which phrases may be decorated on repetition. But to the experienced player, “may be” is not necessarily “must be”’ (128). The second point touches on a broader issue, though Irving does not look beyond Mozart, or indeed the movement under discussion. He observes that while ‘Mozart bothers to overlay his score with a plethora of subtle expressive markings,… the actual sound of this music is something else entirely: it sounds like an improvisation’ (130). Quite true: over-elaborate notation, especially in rhythmic details, often signals a maximum effort on the composer's part to achieve the impression of expressive freedom, or rubato. Perhaps Mozart is not even the most characteristic example of this; several movements of the mature Haydn sonatas (hXVI:42/i, 48/i, 49/ii) or the Fünf Lieder, Op. 3, by Anton Webern might serve as typical examples.

One hardly encounters a serious book on music today that would be content without constant references to the more fashionable authors of literary criticism, designed to make the study appear well informed or ‘up to the mark’. Part of my problem with the first three chapters of Irving's volume is that the long-winded discussions of the ideas of Saussure, Jauss, Gadamer, Iser, Fish and others – valuable in themselves, of course – are not in harmony with the basically practical character of the book. My other problem is that the equally lengthy discourse upon two-part vs three-part sonata form in these same chapters, as I see it, does not lead to any fresh revelations. Surely the textbooks by Koch, Reicha, Galeazzi, Marx and others have been sufficiently treated to detailed accounts in the recent as well as the not so recent literature. And it must be my fault, but I find Irving's idea of an ‘ethical choice’ for the performer between playing an opening movement according to the structural concept of the composer's own lifetime (two-part), and playing it according to that of a subsequent generation (three-part), simply incomprehensible (see the discussion on pages 40–46). Perhaps in a movement like the Allegro of k283, with only a short ‘mid-section’ after the double bar, one might render the old binary form perceptible; but, paradoxically, Irving's model examples for the analysis are movements with exceptionally substantial development sections, like the Andante of k533, or the Allegro maestoso opening movement of k310. How do I ‘communicate’ whether I consider the highly dramatic development section of the A minor sonata as the first half of the second part of a two-part form, or the middle part of a three-part form?

The message of Irving's book is that the ‘understanding of Mozart's piano sonatas’ happens through performance, in a manner that is appropriate to the musical and social context of Mozart's own time, as a reflection of the personal practice of Mozart as composer-performer. To give more weight to his argument, the author stresses the importance of a flexible and quasi-improvisatory approach as something fundamentally different from the ideals of the nineteenth century. ‘Faithfulness to the score is the absolute touchstone of the virtuoso tradition [of the romantic era]’, he states in a footnote in the introductory chapter (‘Pretexts’, page 5), in the course of discussing the art of the nineteenth-century virtuoso musicians. Now, ironically, another recent book by another keyboardist/musicologist about composer-performers of this later period concludes just the opposite. Kenneth Hamilton's highly entertaining as well as scholarly volume After the Golden Age: Romantic Pianism and Modern Performance (New York: Oxford University Press, 2008) demonstrates very similar documents and principles – in a different historical context, of course – to those we read about on the pages of Irving's Mozart book. That is, Hamilton shows the tremendous imagination, freedom and artistic vigour of the great pianists from Liszt to Paderewski, in contrast to the pedantic playing of their twentieth-century successors. (‘Oh, you need not take that so literally’, Franz Liszt remarked to a pupil who strove to reproduce every nuance of the printed score; see Hamilton, After the Golden Age, 228.) The history of musical interpretation is surely a fascinating and complex matter; but might we not perhaps say that healthy freedom of performance did not really cease to exist until the advent of the score-abiding attitude of the twentieth century? (The issue is closely connected with Lydia Goehr's ‘work concept’ (The Imaginary Museum of Musical Works: An Essay in the Philosophy of Music (New York: Oxford University Press, 1992), a frequent point of reference in Irving's prose.) In Goehr's thesis Irving finds reinforcement for his own ideas regarding pre-1800 performances as partially compliant with the score, as against perfectly compliant post-1800 performances. The highly influential philosophy of Lydia Goehr is undeniably useful in a broad sense; nevertheless, it may not be wholly applicable to the virtuoso piano repertoire of the nineteenth century.)

A grave deficiency of the book is the lack of sufficient music examples. They appear in only two sections, in connection with Cipriani Potter's 1848 edition of Mozart's sonatas (69–72) and in the discussion of embellishments (120–127), yet several other important passages cry out for proper illustration. These concern notational details, a crucial aspect of the volume: the bar-by-bar reproduction of the fingering of k310/i in Louis Adam's Méthode de piano du Conservatoire (1805), for example, proceeds in long rows of digits within the prose, with complicated references to the appropriate pitches and rhythms (47–52); the comparison of slurring between autograph and first edition in k330 and k333, meanwhile, involves pages of bewildering verbal descriptions (59–63) where a music example or two would have made all clear.

Unfortunately, the situation is further worsened by Ashgate's very poor job with the editing of the book. The number of blunders is so high that one dares not even consider more professional assistance, such as giving a tighter structure to the content. (For example, there are a number of instances where a particular movement is analysed more than once; although the standpoint might be somewhat different, redundancies are inevitable.) None of us is exempt from oversight, but if an author identifies the Rondeau en Polonaise middle movement of k284 erroneously as ‘minuet and trio’ (82), or mistakes one D major sonata for another (by ‘strong unisono’ opening he must mean k284 and not k311 (77)), it is surely the editor's task to put things right. In addition to the numerous misprints, the list of metronome markings quoted from Potter's 1848 edition of the Mozart sonatas (66–67) includes incorrect rhythmic values as units for the numbers given (k310/iii; k311/ii); whether these have been taken from the source without rectification or have become corrupted in the present print is hard to tell.

The real disgrace of the volume in this respect, however, is the closing Bibliography. In the ‘Primary Musical Sources of Mozart's Sonatas’ the editions of k330 and k333 appear in reverse chronological order; the ‘Selected Modern Editions of Mozart's Sonatas’ strangely includes the new Peters edition of the violin sonatas, but not the Wiener Urtext of the piano sonatas; the ‘Discography’ omits the complete recording of the sonatas on fortepiano by Malcolm Bilson. (The name of Professor Bilson, one of the first exponents of the historical performance of keyboard music of the classical period, and still among its leading representatives, is oddly absent from the pages of Irving's book.)

The gross errors in the ‘General Listing of Works Cited’ are at times hard to believe: Harvard appears as a city, the site of Harvard University Press; Eulenburg is substituted for Norton (!) as the publisher of the renowned English translation of C. P. E. Bach's Versuch; the original title of the same is capitalized in the wrong way throughout the book (‘… Wahre Art …’); James Hepokoski's and Warren Darcy's vast Elements of Sonata Theory (New York: Oxford University Press, 2006) is retitled Element of Sonata Theory, again not only in the Bibliography, but throughout the volume. The total confusion in the listed items of Bilson's publications, where the printed lines are tangled and the author's name reads as ‘Michael Bilson’, makes one wish for a published list of errata which could be appended to the volume.

Lamentable production standards aside, Irving's second treatment of Mozart's keyboard sonatas is highly recommended reading for practical musicians, and pianists in particular. To the initiated, some aspects will be familiar, but there are plenty of ideas here that will be stimulating for teachers and students alike. ‘Happily,’ Irving confesses in his Preface, ‘I have not yet discovered how to reconcile studying and playing Mozart's music. Were I to do so, the enjoyment of trying would have ended, and the provisionality of meaning that inhabits every page of Mozart's sonatas would have been missed’ (ix). Inasmuch as we are all on the way to such a reconciliation, this slender volume might be a useful travelling companion.