I. When we try to define literary balladry George Lyman Kittredge's 1904 definition of the traditional ballad is useful as it describes a range of key features distinct from the literary ballad:

Not only is the author of a ballad invisible, and, so far as the effect which the poem produces on the hearer is concerned, practically non-existent, but the teller of the tale has no rôle in it. Unlike other songs, it does not purport to give utterance to the feelings or the mood of the singer. The first person does not occur at all, except in the speeches of the several characters. Finally, there are no comments or reflections by the narrator. He does not dissect or psychologize. He does not take sides for or against any of the dramatis personae. He merely tells what happened and what people said, and he confines the dialogue to its simplest and most inevitable elements. The story exists for its own sake. If it were possible to conceive a tale as telling itself, without the instrumentality of a conscious speaker, the ballad would be such a tale. [George Lyman Kittredge, Introduction, English and Scottish Popular Ballads, ed. Helen Child Sargent and George Lyman Kittredge (London, 1904) xi]

In contrast, the author of a literary ballad is visible in the organization of the characters, and the narrator, in particular the first person narrator, finely reflects complex states of mind and plays an important role in developing the plot.

The literary ballad is an imitation of the traditional ballad, though it is a task to clarify which aspects of the tradition are imitated. For convenience' sake we classify five types of imitation:

(1) Direct Imitation
This is the most obvious case in which we can be certain of the original ballad on which the imitation is based: for instance, both David Mallet's "Margaret's Ghost" and Thomas Tickell's "Lucy and Colin" are based on "Fair Margaret and Sweet William" (Child 74B); Oliver Goldsmith's "The Hermit, or Edwin and Angelina" (c. 1761) is based on "Gentle Herdsman, Tell to Me" (Percy's Reliques 2, bk. 1, xiv); Robert Burns' "The Carle of Kellyburn Braes" (1794) is based on "The Farmer's Curst Wife" (Child 278A); and Alfred Tennyson's "The Sisters" (1832) is based on "The Twa Sisters" (Child 10C).
(2) Technical (or Formal ) Imitation
The traditional ballad has several narrative characteristics. The 'ballad stanza' consists of a couplet of iambic tetrameter with an interwoven refrain or of four lines of iambic tetrameter and trimeter rhyming a b c b; sometimes a variant of only tetrameter rhyming a b a b or a b c b occurs. When a poet adopts the ballad stanza, we may conclude in principle that he is consciously imitating tradition. Other conventions are repetition and/or the use of refrain, contrast or parallelism, dialogue, abrupt opening, the fragmentation of story, and use of mystical numbers; examples of the above are to be found variously in Percy's "The Friar of Orders Gray" (1765), S. T. Coleridge's "The Rime of the Ancyent Marinere" (1798), Charles Kingsley's "The Three Fishers" (1851), D. G. Rossetti's "Sister Helen" (1851), and A. C. Swinburne's "The King's Daughter" (1866).
(3) The Imitation of Subject Matter
This is the use of the varied subject matter to be found in the ballad tradition, such as historical wars, tragic love, curses, the supernatural world of ghosts and fairies, and metamorphosis, as in Lady Elizabeth Wardlaw's "Hardyknute", M. G. Lewis' "Alonzo the Brave and Fair Imogine" (1796), William Wordsworth's "The Thorn" (1798), Sir Walter Scott's "Alice Brand" (1810), and John Keats' "La Belle Dame sans Merci" (1819). 
(4) Stylistic Imitation
Poets are sometimes attracted to the stylized narrative manner of the traditional ballad, which are anonymous and impersonal in the sense that neither the balladeer, nor the narrator, nor any character, intrudes his or her personality into the narrative, so that the suppression of emotion is most conspicuous. This creates a dispassionate detachment from what happens in the story, as in William Blake's "William Bond" (c. 1803), William Morris's "Two Red Roses across the Moon" (1858), Walter de la Mare's "The Silver Penny" (1902), William Butler Yeats' "Crazy Jane and the Bishop" (1929), and William Plomer's "The Widow's Plot: or, She Got What Was Coming to Her" (1940).
(5) The Imitation of Traditional Ethos
A sense of mutability, playfulness, irony, humour, and parody characterize the spirit and prevalent tone of the traditional ballad; topicality, sentimentality, and didacticism are the key features of the broadside ballad, as imitated in Robert Southey's "The Battle of Blenheim" (1798), Thomas Hood's "Faithless Sally Brown" (1822), W. M. Thackeray's "Little Billee" (1849), Thomas Hardy's "Ah, Are You Digging on My Grave?" (1913), and W. H. Auden's "Miss Gee" (1937).

All imitations draw on the above five elements.

II.  What is a ballad? In answer to this intriguingly complex question W. P. Ker, one of the greatest ballad scholars of the early twentieth century, ventured to say 'In spite of Socrates and his logic' that 'A ballad is The Milldams of Binnorie and Sir Patrick Spens and The Douglas Tragedy and Lord Randal and Child Maurice, and things of that sort.' [W. P. Ker, Form and Style in Poetry (1928; London, 1966) 3.] And various scholars, who are always bound to be troubled by the obligation of definition, have tried to give their own answers, amongst the many, Kittredge in his one-volume abridgement of Child's collection (1904), F. B. Gummere in The Popular Ballad (1907), Louise Pound in Poetic Origins and the Ballad (1921), Gordon Hall Gerould in The Ballad of Tradition (1932), M. J. C. Hodgart in The Ballads (1950), Albert B. Friedman in The Ballad Revival (1961), and David Buchan in The Ballad and the Folk (1972). All of these studies, and briefer introductions to literary terminology such as M. H. Abrams' A Glossary of Literary Terms (5th ed., 1985) and John Peck and Martin Coyle's Literary Terms and Criticism (2nd ed., 1993), contain definitions of the ballad I shall forgo repeating here; but the point common to all is that when critics discuss the ballad they refer to the traditional pieces readily available in Child's The English and Scottish Popular Balladsand in other supplementary collections by his successors. In spite of the range of definitions stemming from the variety of those ballads, commentators can revel in the ballad canon just as Ker did, in so far as they have the accessible resources of a 'definitive' collection to fall back on.

What then is a literary ballad? Malcolm Laws in his comprehensive and still indispensably informative The British Literary Ballad: A Study in Poetic Imitation (1972) responds to this fundamental question, which all ballad scholars set themselves, with a typical, classic definition, that literary ballads are 'the product and possession not of the common people of village or city but of sophisticated poets writing for literate audiences. They are printed poems rather than songs, and they have no traditional life. Despite great variations among individual examples, the literary ballads as a class are conscious and deliberate imitations of folk and broadside ballads.' [G. Malcolm Laws, Jr., The British Literary Ballad: A Study in Poetic Imitation (Carbondale, 1972) xi] Then, he goes on to say, following Ker, that 'In the field of balladry, definition by example has often been found more enlightening than abstract verbalizing. Thus one may begin by identifying as literary ballads such frequently anthologized poems as the following: Wordsworth's "Lucy Gray," Scott's "The Eve of St. John," Southey's "The Battle of Blenheim," Tennyson's "The Charge of the Light Brigade," Rossetti's "Sister Helen," Housman's "Is My Team Ploughing?", Hardy's "Ah, Are You Digging on My Grave?" and Yeats's "The Ballad of Father Gilligan"' (Laws, Jr. 1). But in the case of literary balladry, the remarkably small number of anthologies published to date is not sufficiently reliable or exhaustive to serve as the basis for any binding definition of the form. As such, Laws' list of literary ballads, though it is a useful orientation, is not, finally, definitive as, arguably, many major, certainly interesting pieces are omitted because, as he concedes, they do not fit his personal definition of balladry (cf. Laws, Jr. 149-61). The ongoing limitations of Laws' principles (which essentially replicate Ker) place the reader at a loss when confronted with a poem, plausibly a literary ballad, which is inexplicably absent from the Laws canon. This highlights the shortcomings of definition by example, which cannot provide a trustworthy gauge for whether a particular old or new poem can be classified as a literary ballad or not.

The essential difficulty of studying literary ballads comes from the very fact that they are the products of sophisticated poets. They are poems in their own right as well as being ballad imitations. The relative emphasis given these two aspects of the form depends upon which aspect of balladry interests an individual poet the most. Yet no matter how much a poet's interest gravitates towards imitation, he can never annihilate himself into the anonymity of the traditional balladeer. Child, as deeply as any other scholar after him, realised this crucial problem:

The condition of society in which a truly national or popular poetry appears explains the character of such poetry. It is a condition in which the people are not divided by political organization and book-culture into markedly distinct classes, in which consequently there is such community of ideas and feelings that the whole people form an individual. Such poetry, accordingly, while it is in its essence an expression of our common human nature, and so of universal and indestructible interest, will in each case be differenced by circumstances and idiosyncrasy. On the other hand, it will always be an expression of the mind and heart of the people as an individual, and never of the personality of individual men. The fundamental characteristic of popular ballads is therefore the absence of subjectivity and of self-consciousness. Though they do not 'write themselves,' as William Grimm has said, though a man and not a people has composed them, still the author counts for nothing, and it is not by mere accident, but with the best reason, that they have come down to us anonymous. Hence, too, they are extremely difficult to imitate by the highly civilized modern man, and most of the attempts to reproduce this kind of poetry have been ridiculous failures. [Quoted in Walter Morris Hart, "Professor Child and the Ballad", PMLA 21 (1906): 756-57]

The justice of the charge made by Child of the imitations' failure must be debated, but may be countered to a large extent by the enduring interest the ballads have held for poets historically. The balladeers' envied relation to art form and audience has enjoyed a repeated privilege of discussion, but is reiterated most tellingly in the act of imitation itself, where a poet's urgency for anonymity contends with the overall self-consciousness of the performance. To return to T. S. Eliot's celebrated observation that 'poetry is not the expression of personality, but an escape from personality' ["Tradition and the Individual Talent", The Sacred Wood (1920; London, 1960) 58] is to be close to the general constitution of the literary balladist, whose art offers the poignant drama of the poetic self resisting oblivion.

(Partly abstracted from Mitsuyoshi Yamanaka, Introduction, The Twilight of the British Literary Ballad in the Eighteenth Century, Kyushu UP, 2001)