342 PART V. WORKING STAGE. 



Com jnence 

 Co n nection 

 Co - operation 

 Complexion 

 Co n quer 

 Co r rection 

 Co n slant 

 Co n /inue 

 Co "- Unitarian 

 Co n yiction 

 Co w 

 Co x 

 Co y 

 Co z 



Everything becomes now clear as day, and without delay I am 

 in the position to formulate the ensuing apposite conclusions: 



(a) No con-combinations exist with non-Latin words, L e., 

 words beginning with k, w, x, y, z. 



(b) Stems whose first letter consists of a vowel or the aspirate 

 are preceded by co, though exceptions exist. 



(c) In certain cases where the n in con-combinations would 

 be difficult to pronounce, it becomes 772, as in words commencing 

 with b, p, m. 



(d) For similar reasons as in (c), the n in con becomes / be- 

 fore /, and r before /. 



(e) Rules (a), (c), and (d), apply presumably to in and other 

 prefixes terminating in /?, if others should exist imbibe, im- 

 pinge, z/legible, immaterial, zrreductible, etc. 



Similar rules are presumably operative throughout the English 

 language in the words derived from Latin, e.g., rampant, e/fusion, 

 diffusion; or, ad adapted as follows: a/firm, aggrandise, a/lure, 

 ammunition, apparent, arrive, ascent, assume, attention. 



From this lengthy statement it will be readily inferred that 

 habitual and methodical procedure as a whole, including of 

 course habitual and methodical generalisation, are of prime im- 

 portance. 



The sub-headings in Conclusion 16, pertaining to Observation, 

 should be also utilised in connection with the process of gene- 

 ralisation. 



CONCLUSION 26. 



Need of Postponing large Generalisations to near the Conclusion 



of the Enquiry. 



173. Useful as is the formulation of comprehensive con- 

 clusions at the termination of an investigation, it is mischievous 

 to seek for these at the commencement, for the reason that 

 the conclusions then arrived at are almost certainly premature 

 and are likely therefore to be erroneous and misleading. Until 

 the field under inspection is scrupulously explored, the largest 

 number of conclusions are collected, and the ground is traversed 

 repeatedly, it is well to regard the conclusions reached as 



